Training Coach
Doctrine
A report may cite any of these as doctrine-N.
doctrine-1Contract freeze and fingerprint disciplineThe policy I/O contract (observation layout, scales, history semantics, action pipeline) is frozen and fingerprinted; every exported policy is stamped and verified; contract changes ship as new versioned profiles that leave old artifacts bit-identical, and old policies run forever under their era's pinned profile.
Case. The 215-dim omni contract was frozen with a three-machine digest; the one contract-level extension (lateral feed-forward) went in as a new `omni_ff` profile with the old profile provably untouched, and the contract checker caught two real wiring bugs before any training (`contract-freeze-and-checker`). A silently changed gait-clock default would have fed old policies a 25% slower clock - closed by pinned legacy profiles (`legacy-profile-pinning`). A stale derived USD forked plant mass 2.2% until an automated source-vs-derived instrument gated it (`derived-asset-staleness-check`). A gain profile is part of the closed loop a policy was trained in and belongs in its stamp; the recovery line's anchored authority was left out of its manifest and recorded as the gap not to repeat (`gain-profile-belongs-in-the-stamp`), and a second policy behind a deploy-side switch made the handoff state itself a contract (`recovery-two-policies-and-a-state-machine`, `walk-recovery-fsm-handoff`).
Coach application. On any proposal touching obs/action semantics, defaults, or derived assets: demand the version/profile plan, the fingerprint update, and the checker extension in the same change; flag any old artifact that would run under new defaults.
doctrine-2Attribution by resolved training params - never eval-override knobsCapability differences between lineages are explained only by digging each lineage's *resolved* training configuration and eliminating columns; evaluation-side override knobs (kd-scale, power-scale, cycle-time) act on the plant for *every* policy and may serve as deployment mitigations but never as explanations.
Case. Low-friction robustness across 8 lineages x 3840 cells was traced to kd DR *bandwidth* - every lineage had ground friction pinned to (1.0,1.0), so "trained friction" could not be the axis; the parameter axis and the plant axis were explicitly separated after the first attribution conflated them (`kd-bandwidth-mu-law-attribution`). "Weak turning" on hardware was a power-scale plant effect, not a training gap (`deploy-knob-attribution-before-retraining`); slowing the deploy clock was out-of-distribution, not a feature (`cycle-time-override-is-ood`). The ground truth for what a run trained under is the logged per-run config, not the source tree (`resolved-config-is-source-of-truth`).
Coach application. Whenever asked "why is lineage A better", require the resolved-param table first; kill zero-variance columns; refuse explanations phrased in eval-knob terms; when a knob helps, label it deployment mitigation.
doctrine-3PASS gates become constraints; FAIL gates become objectivesOnce a skill passes its gate, that gate converts into a standing regression constraint (budget <= 2/20 against the parent baseline) for all later training; gates currently failing are the only legitimate objectives of the next rung.
Case. The C ladder ran one frozen 13-cell x 20-seed matrix at every rung with promotion = "new skill PASS and old skills within regression budget"; C1 was stopped and re-rooted precisely because it trained away the root's backward PASS (`fixed-acceptance-matrix-per-rung`, `preregistered-stop-criteria-per-rung`). The C4 product shipped only at 260/260 cells with zero regression.
Coach application. Keep the ledger: every PASS adds a constraint row; propose rungs only against FAIL rows; treat any constraint violation as stop-and-attribute, never "the next rung might win it back".
doctrine-4One variable per ladder rung - counted against what the checkpoint sawA rung changes one variable, where "one" is counted against the checkpoint's actual training state, not against the current config's diff; batching is allowed only when each change owns a disjoint symptom space with a pre-registered ablation order.
Case. Two rungs failed identically because resuming s1e-500 under the evolved config silently added four plant variables the checkpoint had never seen ("单变量纪律不只看「我改了什么」,还要看「checkpoint 见过什么」" - `resume-state-dr-audit`). v8 legally batched four orthogonal fixes with a written ablation order (`orthogonal-batch-with-ablation-order`); v9 spent one run completing a 2x2 factorial so either outcome convicted a factor (`fill-the-missing-factorial-cell`); v10b's three-way ablation wrongfully convicted the clock and had to be retried fairly.
Coach application. Before any resume: diff cfg against the checkpoint's logged training state. Before any batch: require the symptom-ownership map and ablation order in writing.
doctrine-5Pre-register risks, readings, and stop criteria before the ladderBefore a ladder or risky rung, write down the known risks, the interpretation of every plausible outcome, and hit-any-one stop criteria - frozen before training, tightened when priors say results should come fast.
Case. The C ladder opened with three numbered risks including the exact falsification condition for its own root choice; A/B arms carried "预注册读法(事后不改)" tables; a level expected to fail was run anyway for its pre-registered diagnostic value (`preregister-risks-and-fork-readings`). Stop criteria caught C4-redo rungs at +200 instead of full caps (`preregistered-stop-criteria-per-rung`); hardware sessions pre-registered per-config expected signatures and the disagreement rule "不改结论改账" (`preregistered-real-expectations`, `feasibility-accounts-lock-design-point`).
Coach application. Refuse to open a rung without the written risk/reading/ stop block; after results, read conclusions off the pre-registered table and flag any post-hoc reinterpretation.
doctrine-6Plant parameters are measured, never inventedEvery plant number carries measurement provenance: armature = N^2 x rotor inertia from no-load tests, friction split by rig and by API column, torque limits shaped by per-joint gait peaks, latency traced through the real pipeline, masses weighed - and DR bands are additive around the measured nominal, sized to the measured dispersion.
Case. Guessed friction was 2.5x low and guessed damping 5x high (`friction-measured-not-guessed`); armature had been 0 with a 9:1 gearbox (81x reflected inertia, `armature-n2-rotor-inertia`); a uniform torque derating was "the wrong shape" vs measured peaks (`torque-limit-shape-by-measured-peaks`); the delay implementation itself was a wrong plant for a whole lineage (`latency-lerp-reverse-extrapolation`); the run design point was locked by three accounts including the tau_limit/kd speed ceiling (`feasibility-accounts-lock-design-point`); identified friction had to land in the right simulator API columns to act at all (`sim-api-friction-columns`). The recovery and one-leg lines opened with the same kind of accounts before any reward existed - a connected static path and the torque along it for an armless get-up, and the gains single support needs to be holdable at all (`get-up-feasibility-accounts-before-training`, `single-support-gain-authority-probe`).
Coach application. For any plant value in a config review, ask "measured how?"; reject absolute ranges with no nominal; check API column mapping and derived-asset regeneration whenever measured values land.
doctrine-7Sim2sim gate before sim2real - under deployment conditionsEvery checkpoint passes a second, independently built simulator before hardware, and both the gate and the smoke loop run under the measured deployment conditions (real pipeline delay, honest contact parameters, the deployment gain/power profile).
Case. The standing order "先sim2sim 再sim2real" (`sim2sim-gate-before-sim2real`); acceptance flipped to match hardware only under measured condim/torsional friction (`eval-plant-honesty-contact-params`); gates moved permanently to `--delay 2` after the kicking incident (`pipeline-latency-is-plant-not-dr`); and the harness itself must be audited - a frame-convention bug in the cross-sim evaluator invalidated a whole line of verdicts (`body-frame-velocity-api-audit`). The recovery line's second simulator caught a torque penalty paid for by bracing the legs together (`torque-penalty-bought-by-leg-bracing`), and a 1.8x torque disagreement between the two plants stayed binding because its one surviving explanation was never tested (`torque-disagreement-between-simulators-unresolved`).
Coach application. Block any hardware request lacking a second-sim PASS at deployment conditions; when sim2sim and training-side metrics disagree, treat the evaluator as a suspect too.
doctrine-8Observation honesty - the actor's inputs are a hardware contractThe actor observes only signals the real robot produces with realistic noise; privileged truths go to the critic; history windows are estimators and must train under plant variation; rewards on quantities the actor cannot observe buy only average suppression, never closed-loop correction.
Case. Ground-truth velocity/forces went critic-only (`observation-honesty-critic-only`); frame_hist under zero DR memorized the trainer's plant fingerprint - 0/20 transfer (`history-obs-needs-plant-variation`); world-frame yaw rewards could not teach pull-back because heading is unobservable to the actor - correction was routed to the deploy outer loop instead of breaking the contract (`reward-observability-limit`, `deploy-heading-loop-and-align-training`).
Coach application. Audit every actor-obs element for hardware existence; require minimal plant jitter whenever history/recurrence exists; for each reward, ask "can the actor see this error?" and route correction tasks to outer loops.
doctrine-9Reward economics are audited in realized currencyReward design decisions are made on realized per-step magnitudes under the actual policy and command distribution: price the do-nothing optimum before adding a mode, compare achieved values to the computed ignore-floor, calibrate thresholds between measured healthy and sick distributions, and ship every new penalty with a withdrawal clause.
Case. feet_air_time at weight 2.0 realized 0.038 vs tracking 1.2 - drag was rational (`realized-contribution-audit`); ignoring a vy command cost 28-180x less than ignoring vx until a gated tracking term was added (`reward-cost-of-ignoring-audit`, `gate-new-reward-terms-by-command`); achieved-vs-floor separated "never learned" from "priced out" (`ignore-floor-diagnosis`); the foot-distance wall was placed between measured healthy (0.6% tax) and sick (55%) policies (`calibrate-threshold-between-healthy-and-sick`); the landing penalty carried a pre-registered stand-down condition and actually stood down (`calibration-threshold-with-withdrawal-clause`); two clearance terms were inert until zero-points and gate occupancy were checked (`inert-reward-term-audit`). A get-up policy sat because three gated terms paid the seated pose 84% of the return and the one term that could tell sitting from standing was an exp kernel reading 4.6e-5 at the real error (`seated-basin-dead-exp-kernel`); a torque-tail term was weighted by its measured steady value beside a peer term after the estimate proved 12x off (`tail-torque-needs-hinge-on-computed-demand`).
Coach application. Never discuss weights in the abstract: demand the realized-contribution table, the ignore-floor number, and the healthy-pay calibration before any reward edit is approved.
doctrine-10The zero-cost option must be the desired behaviorFor every penalty, name what the zero-cost option is; penalize failure events (slip, saturation excess, contact in flight windows), never the motion or joints that healthy behavior uses; make degenerate strategies fatal via termination where penalties cannot price them out.
Case. Joint-usage penalties for drift taxed a 1.4%-of-momentum channel 2.7/step and collapsed training; the slip penalty costs a non-slipping gait exactly zero (`penalize-the-slip-not-the-joint`). A frozen-at-clamp joint pays zero action-rate forever - only a pre-clip saturation penalty flips the cheat economics (`saturation-cheating-zero-rate-cost`). Ungated phase shaping made standing 42x more expensive than stepping and cooked the hip motors (`moving-gate-42x-stand-tax`); crouch-shuffling lived until a height termination deleted it (`termination-closes-degenerate-basin`). A gated penalty is an exit: the policy parked just outside an uprightness gate, then just under a height gate, to stop paying a stance tax, and only a positive band plus an always-on guard closed both (`penalty-gate-is-an-escape-hatch`); a soft-limit penalty that charged the standing pose itself bought a 4.1 deg lean (`soft-limit-penalty-charges-nominal-pose`); an unpriced foot attitude was spent on edge-standing (`unpriced-foot-attitude-is-a-free-variable`); and the one-leg line listed its cheapest cheats before training and still met one through a zero-gradient band (`enumerate-cheapest-cheats-before-training`, `binary-band-reward-fake-touchdown`).
Coach application. Run the "零代价的选项是什么" audit on every proposed term; convert motion taxes into event-conditional penalties; check the termination set against each known degenerate strategy.
doctrine-11Measurement discipline: independent referees, signs, distributionsA disputed measurement is adjudicated only by an independent algorithm from raw state; directional ability requires sign-antisymmetry under command reversal; bimodal metrics are reported as mode shares (never medians, never 3 seeds); ratios are not comparable when totals change; reward values compare only within one command distribution; single chaotic events never cross machines.
Case. The triple reversal - a good metric was "refuted" by a sibling metric that shared the disease (`independent-referee-for-metric-disputes`, `body-frame-velocity-api-audit`); same-signed +/- responses were bias, not turning (`same-sign-response-is-yaw-bias`); the swing median sat in a bimodal gap (`median-hides-bimodal-distribution`); "v6 is jitterier" died on absolute energies (`ratio-metrics-need-absolute-check`); yaw gain measured 15x wrong in an oscillating frame (`heading-integral-not-body-rate`); a 44% improvement evaporated under same-distribution comparison (`same-distribution-reward-comparison`); drift direction was a limit cycle (`multiseed-sign-test-for-drift`); a cross-machine push cliff was chaos (`single-impulse-recovery-is-chaotic`).
Coach application. Before accepting any surprising number: ask for the independent recomputation, the sign pair, the distribution shape, and the comparison conditions. Retract in writing when a metric falls.
doctrine-12The deployment pipeline is plantIrreducible pipeline properties - action latency, rate limits, power/torque scaling, teleop command mappings - are part of the nominal plant, modeled from day one and reproduced in every gate; deploy-side scalings are crutches that flag unmodeled plant, and they cannot be algebraically folded into training constants.
Case. Right-leg kicking was over-trained-delay x loop gain; power 0.8 was a gain-reduction crutch that retired when the delay was modeled (`pipeline-latency-is-plant-not-dr`); power derating damages non-forward axes first (`power-scale-hurts-nonforward-axes`); training at 0.4 scale as the "twin" of deploying 0.5 x 0.8 collapsed 0/20 (`deploy-scaling-not-training-equivalent`); one shared teleop speed sent an out-of-band lateral command and the robot clipped its own foot (`teleop-command-band-per-axis`); the latency DR range had not even covered the measured pipeline (`latency-dr-covers-measured-pipeline`). A rate limiter added at deployment only clipped a policy that kept commanding (`deploy-rate-limiter-windup`); moved into training and anchored on the last command it became an integrator in the balance loop (`slew-anchor-is-an-integrator`); anchored on the measured angle it bounded torque and kept the bandwidth (`beta-anchored-action-target`). The walking lines' safe setting, power-scale 0.8, cut the ends of the recovery policy's full-range travel and left its spikes alone; a gain inside the trained band did the job (`power-derating-cuts-full-range-contract`).
Coach application. Demand the measured pipeline latency/limits in the plant model and in gate conditions; treat every deploy-side derating as a question ("what is this compensating?"); block per-axis command sources that exceed training bands.
doctrine-13DR budget is finite; its distribution is the measured supportRobustness is a conserved budget: disturbance training on an already-hardened lineage borrows from existing margins; DR ranges span the measured deployment support - no fictitious tails (they buy degenerate gaits), no single constants (they allow thin-margin specialization); harden the plant only after the task distribution is final.
Case. The same push dose helped a narrow lineage and damaged a balanced one - budget conservation (`push-dr-conditional-budget-conservation`); wide latency tails bought drag-glide, constant values shipped 60% thinner tilt margins - the answer is a narrow band on the measured support (`dr-tail-plant-continuation`, `constant-value-dr-overfits-margin`); task-first ordering because hardening a soon-to-change task wastes budget (`task-shaping-before-plant-hardening`); COM randomization used deliberately as a behavior-shaping tool, and rolled back on symptom per its own contract (`com-randomization-forces-leg-spread`, `com-dr-rollback-on-symptom`). DR that is switched on can still be thin: the run policy fell in the frontal plane its gain-and-latency randomization never touched (`thin-dr-judged-by-channel-coverage`), and a friction priority settled under one action contract had to be re-measured under the next (`friction-priority-re-measured-after-plant-change`).
Coach application. Before any DR rung: check the untrained policy against the spec, the lineage's current DR load, and the measured real-world range; after it: audit retained margins, not just the new tolerance.
doctrine-14Gates measure what hardware feels: posture, margins, stripped assistsAcceptance batteries carry posture-class rows (tilt max median, per-joint L/R asymmetry, temperature) beside task rows, graded margin columns beside binary gates, chirality scored per side, at least one condition that removes the environment's free stabilization, and validated predictive scalars promoted into the gate.
Case. Three same-shaped judging errors - survival, displacement, wz-difference - all missed what the operator felt; posture metrics had the predictive power (`task-metrics-vs-posture-metrics`, `stand-gate-posture-not-survival`); binary survival saturated and hid a 60% margin gap (`constant-value-dr-overfits-margin`); v5 passed everything on the ground and failed suspended (`suspension-probe-removes-free-stabilizer`); the hip_roll (l+r) scalar predicted real drift direction and ordering and entered the battery (`hip-roll-sum-predicts-lateral-drift`); averages hide chirality (`chirality-scored-separately`); gait-quality gates are judged at speeds that demand a gait (`low-speed-commands-reward-dragging`). The recovery line added the rest of the kit: where failed episodes end, not only where they started (`end-state-confusion-matrix`); a frozen acceptance distribution with a pinned seed (`frozen-acceptance-distribution-and-pinned-seed`); video of the metric rollout itself (`video-as-acceptance-record`); and the admission that a 10 s episode cannot see a stance that fails after a minute (`episode-length-bounds-what-a-gate-sees`). The one-leg line removed a foot-spacing wall that no gate measured, and the feet met on hardware (`removed-wall-returns-on-hardware`).
Coach application. Review every battery for posture rows, margin columns, per-side scoring, and an assist-stripped condition; when operator feel and gates disagree, suspect the metric class first.
doctrine-15Fork and root selection: recoverability, maturity, frozen rewardsChoose fork roots by which candidate's deficits the coming training can pay back (precision is recoverable; lost plasticity, symmetry, and margins are not); prefer mature checkpoints as roots even when younger ones score better as products; never fine-tune through a reward change - continuation is legal only with the reward frozen and plant/DR widening one rung at a time.
Case. s1e-500 beat higher-precision candidates because its exclusive strengths were unrecoverable (`fork-root-recoverable-shortfall`); the b300 arm proved maturity is capital against adaptation shock (`root-maturity-vs-product-quality`); the B-arm scatter/half-recover/collapse signature falsified reward-change fine-tuning and drew the legal boundary for S2 continuation (`fine-tune-reward-change-falsified`).
Coach application. For root debates, build the exclusive-strengths table and ask "which side can be trained back?"; require dual-arm evidence for maturity claims; classify any proposed continuation as reward-frozen or not before approving.
doctrine-16Curricula: verified engagement, lineage counters, disease-phase gatingAutomatic curricula must prove they engage (a saturated ratchet is constant DR wearing a curriculum's name); every ramp counts lineage-cumulative progress, not per-process steps; penalties aimed at late-stage pathologies ramp in after exploration noise decays; difficulty rises on measured per-stratum success, never on schedule.
Case. The s1f ratchet capped at iter 248 and never engaged (`auto-curriculum-engagement-check`); the saturation ramp re-fired at +600 after every resume and no shipped product ever saw the penalty (`curriculum-counter-lineage-steps`); the same penalty worked once gated to the disease phase and became an untouchable mechanism (`gate-penalties-to-the-disease-phase`); record-high aggregate reward hid a fully-failing delay stratum (`aggregate-metrics-mask-subgroup-failure`); bucket share is not a gradient lever (`bucket-share-is-not-a-gradient-lever`). An assist curriculum keyed to a pooled success share was withdrawn on the strength of the categories that already worked (`curriculum-criterion-conditioned-on-lagging-category`); a pace set by per-step income moved only when that income was time-gated (`per-step-income-drives-speed-time-gate`), and the same gate had to be retired in a lineage without the disease (`time-gate-vs-wide-stance-retire-the-fix`).
Coach application. Ask every curriculum three questions: does it engage (show the internal state)? what does it count (process or lineage)? when is it present (against the pathology's phase)? Check where shipped checkpoints sit relative to every ramp.
doctrine-17Probe before training: feasibility first, hypotheses in tablesAfter two failed training attempts at a skill, stop training: demonstrate the behavior open-loop, enumerate hypotheses in a written table audited against actual configs cheapest-first, race one probe per side of the sim2real boundary for hardware-only pathologies, and use suspended tests to acquit or convict actuators before blaming authority.
Case. "在黑暗里试钥匙" - four sidewalk rungs failed until an open-loop probe separated exploration/waveform/authority in one experiment (`open-loop-probe-before-reward-tuning`); the foot-drag mystery fell to a seven-hypothesis config audit (`hypothesis-table-code-audit`); the period-doubling was resolved by racing a reward-side and a plant-side evidence line - and both paid off, one per sub-case (`period-doubling-evidence-race`); the suspended test acquitted the roll actuator in one measurement (`suspended-test-isolates-actuator-authority`). A read-only configuration probe told a wall from a slope in the recovery line's seated basin (`configuration-probe-wall-not-slope`), and the fix it pointed to - where the feet are - took prone from 0/159 to 158/159 (`prone-dead-end-is-foot-placement`); a knob that did not move its variable was recorded as no test of the idea (`dof-vel-penalty-is-not-a-pacing-knob`).
Coach application. When a skill resists training, prescribe the probe before any further reward edits; require verified target trajectories before imitation terms; keep a falsified-fixes list so closed roads stay closed (`amplitude-cut-falsified-yaw-fix`).
doctrine-18External advice is recomputed locally; values transfer as ratiosEvery external suggestion is classified adopt / already-have / modify / trap by recomputing its claim on the local reward table and probe data; numeric values transfer only as dimensionless ratios (to tracking weight, leg length, sqrt(gL), control rate); citations are verified to exist.
Case. "Start vy very small" would have destroyed sidewalk learning on this reward table - the gradient scales quadratically (`external-advice-audit-against-own-arithmetic`); swing-height targets and weights transferred correctly only through leg-length and tracking-ratio scaling (`transfer-ratios-not-absolutes`); the "6-step delay" was refused for lacking a control rate (`latency-dr-covers-measured-pipeline`); a borrowed reference's structure was FK-verified and its amplitude re-derived from the division of labor (`reference-structure-fk-amplitude-division`); retrieval agents fabricated verbatim arXiv quotes - only source-verifiable material was used; and one dismissed suggestion later proved right for a different mechanism, and was credited (`cycle-average-tracking-for-gait-quantities`). An advisor's staged state machine turned out to exist in none of the three papers it cited, and reading them changed the plan (`advisor-paraphrase-vs-paper`).
Coach application. Intercept every "paper X does Y" with the local recomputation; convert absolutes to ratios before comparison; verify quotes; revisit dismissed advice when new mechanisms appear.
doctrine-19Hardware sessions are scripted experiments, not tuning sessionsReal-robot time executes a pre-registered matrix: risk-ordered (baseline first, fragile last with a spotter), stage-gated (suspended smoke before ground), A/B sessions bracketed by a repeated reference run, operators briefed on measured zero-command and untrained-axis behavior, chirality-aware disturbance protocols, no field tuning - the only legal field changes are scripted, single-variable, and self-reversing.
Case. The S2 acceptance sheet (`risk-ordered-real-deployment`, `battery-bracketed-real-ab`, `know-zero-command-behavior`, `push-test-chirality-protocol`, `no-field-tuning-protocol`); the RAM-only torque experiment with automatic power-cycle rollback (`reversible-single-variable-field-experiments`); and the sim-veto rule - even sim's condemnations get one safeguarded hardware check when they judge the purpose-built configuration (`sim-veto-needs-real-confirmation`). The recovery line's first real run went ahead with its preconditions unmet and was stopped as dangerous (`first-real-get-up-violent-stage-one-policy`); after it: a staged hang, mat and floor protocol (`staged-hang-mat-floor-for-get-up`), a fixed power-cycle pre-flight and two-machine discipline (`power-cycle-preflight`, `two-machine-config-discipline`), a fall guard replaced rather than switched off (`fall-guard-becomes-a-state`), and logs that are part of the run (`hardware-log-is-the-attribution-input`).
Coach application. Turn every hardware request into a runbook with order, gates, brackets, briefing, and anomaly plays; refuse improvised parameter changes on the floor.
doctrine-20Close questions in writing; restart when the debt is structuralAudited questions get frozen verdicts with citable wording and an explicit reopening bar; hardware verdicts are dated by deployment-stack and calibration state and expire when those change; and when successive rungs shuffle symptoms without net progress, freeze the lineage as regression baselines, pay the structural debts, and retrain minimal - carrying laws and instruments, not weights.
Case. The chirality and COM questions were closed with frozen wording and "no reopening without new hard evidence" (`frozen-verdicts-semantic-boundaries`); v5/v6's condemnations expired with the deploy stack (`stale-verdicts-under-old-stack`); a 2-degree calibration fix moved the whole runnable envelope (`zero-offset-calibration-shifts-envelope`); plant upgrades are era boundaries with paired re-baselining (`plant-swap-invariants-vs-shifts`); and the 2026-08-05 reset froze v5-v11, fixed the latency FIFO / manifest / sampling / reward-table debts, and restarted - producing the lineage that reached hardware SOTA (`freeze-lineage-fix-structure-restart`, `minimal-reward-table-with-provenance`). The recovery line's real-robot verdicts ended up in three places that disagree, one of them an undated note in a command file (`write-hardware-verdicts-back`).
Coach application. Maintain the closed-questions ledger and quote it when symptoms recur; stamp verdicts with stack/calibration versions; when a team is three rungs into symptom-shuffling, raise the restart question explicitly with the freeze-fix-restart pattern.
doctrine-21Name the quantity in the space it lives inA goal, reward term or acceptance criterion about the feet, the base or the contact state is computed from the quantity itself - world poses, forces, per-category outcomes - never through a joint-angle, single-signal or pooled stand-in that assumes everything else sits at nominal; and every detector is validated on a behaviour known not to contain the event before it becomes a gate.
Case. The recovery line was caught three times: |ankle roll| as "flat feet" sold stance width and the real robot slid into the splits, a hip-roll criterion was confounded by 50 deg of yaw, and the joint table said 0.271 m where the feet were 0.159 m apart; task-space terms produced the first flat, wide stance (`joint-space-proxy-for-task-space-quantity`). Flight detection lied in both directions across two lines - foot height flagged 40% false flight on a walking gait, contact force alone flagged slip chatter as hops (`contact-detector-single-signal-lies`). A pooled height average described a robot that did not exist - six in ten standing, four in ten sitting (`zero-partial-credit-is-not-an-iteration-problem`) - and the walking line had learned the same lesson on yaw rate (`heading-integral-not-body-rate`).
Coach application. For every reward term and gate row, ask what physical quantity it stands for and whether it is measured directly; flag joint-space or single-signal stand-ins for task-space goals, ask for a detector validated on a negative control, and split pooled metrics by category before reading them.
doctrine-22Continuation needs a live gradient; a release is chosen by a scanContinue a converged policy only on a change that creates a live gradient, on a short budget, with every checkpoint scanned on the transfer axis; choose a release by running the full battery over a band of checkpoints and stop on signals, never by taking the last one; and when edits to the terminal phase cannot move a behaviour, roll back and retrain with the constraint present from the start, keeping the order in which the lineage acquired its mechanisms as explicit curriculum phases.
Case. A continuation with no new gradient drifted MuJoCo transfer from 100/98% to 80/28% while every Isaac gate stayed perfect, and a live-gradient continuation at the same depth kept it (`converged-continuation-is-poison`). One-leg checkpoints 100 iterations apart failed 1 and 38 of 40 cells, and late ones degraded (`checkpoint-choice-is-a-full-gate-scan`). Four in-lineage stance fixes failed because the stance was the end of the get-up path, and from scratch it grew right (`stance-decided-by-get-up-path`); fixes stacked on degraded states were rolled back by the user (`stop-stacking-roll-back-and-audit`); and the lineage's final recipe, trained from scratch in one run, sat at 0% because the order of its curriculum was part of the product (`curriculum-history-is-part-of-the-product`). The omni line's short adaptation budgets and mature roots are the same law seen from the other side (`continuation-budget-not-from-zero`, `root-maturity-vs-product-quality`).
Coach application. Before approving a continuation, ask for the new gradient, the budget and the transfer axis in the scan; before approving a release, ask for the scan; after three rungs without progress on the target, propose rolling back to the last good checkpoint and a from-scratch phase plan instead of a fourth patch.
Experience cards
171 cards matching “curriculum-history-is-part-of-the-product”.
Training the final recipe from scratch in one run - every mechanism the lineage had accumulated - produced 0% and a seated robot; the order in which the lineage acquired those mechanisms was part of why it worked
curriculum-history-is-part-of-the-productA recipe that ends a lineage is not a recipe for a from-scratch run: consolidate it as ordered curriculum phases matching how the lineage acquired its mechanisms, check that every curriculum criterion is reachable from the starting policy, and read the run's raw term values, not the total reward, before calling it green.
Symptom
V3.0 trained the lineage's whole final recipe from scratch in one 9,000 iteration run - full beta curriculum, prone-conditioned pull assist, friction DR, the 3 s zero gate on standing income, flat_feet - testing the proposition "the product is defined by its configuration, not by its training history". Training looked all green (reward 26.33, episode length 500, 100% time-outs).
Context
Read in raw units against v2_6c at the same weights, the green was a seated equilibrium: base_height 0.377 vs 0.640, stand_pose 0.205 vs 0.516, flat_feet 0.0000 (zero because it sits outside its height gate, not because the feet were flat). Acceptance: 0.0% in Isaac at the deployed authority, 0.0% on every MuJoCo friction level, and still 0.0% at the training-time authority (100% seated at 0.222 m, upright and still).
Change
The full stdout (316k lines) was read: the beta curriculum's criterion (standing share over 0.35) was met zero times, so beta never left the wide setting and the policy had no experience at the deployed authority; the pull curriculum was stuck on the same criterion. The lineage had escaped the seated basin with immediate income (V2.0-V2.2) and only then added the zero gate to cure rushing (V2.5); from scratch, the zero gate removed the early "stand fast, earn more" gradient needed to escape. Verdict "the curriculum history is part of the product", limited to n = 1. v2_6c stayed the product.
Outcome
V3.1 kept the order as explicit phases: P1 from scratch with immediate income (zero gate off) until the curricula advance, P2 adding the zero gate. P1b/P1c escaped the seated basin and passed; P2 was later judged net negative and dropped (time-gate-vs-wide-stance-retire-the-fix).
Mechanism
Mechanisms that refine a competent policy (time gates, tight authority) can delete the gradient a naive policy needs, and a curriculum whose advancement criterion the naive policy never meets freezes at its first level.
Conflicts
The spec limits the falsification to "this recipe + this curriculum criterion" (n = 1, no seed sweep, no criterion tuning). V3.1's phased run succeeding is consistent with the ordering reading but changed other terms too.
Applies when
- consolidating a long lineage of continuation fixes into one clean recipe
- a from-scratch run with all mechanisms enabled plateaus early
- curriculum state is not logged or never advances
“命题:产物由配置定义,而非训练史定义。 … 训练 log(完整 stdout 316k 行)`[beta_anchor]` 仅初始 1 行,**达标 0 次** … V2.0~V2.2 靠**即时计酬**爬出坐姿盆地(§33),站立巩固后 V2.5 才装归零门 治"过快"(§39)。**课程史是产品的一部分。**”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §42 V3.0 判决(2026-08-11):从零单 run 全机制 FAIL 于坐姿盆地 An auto-curriculum ratchet capped out at iter 248 and never engaged - stage difficulty manually or verify engagement
auto-curriculum-engagement-checkPrefer manually staged difficulty with gated transitions; if you use an automatic curriculum, instrument its internal state and alarm when it stops engaging - a saturated curriculum is constant DR wearing a curriculum's name.
Symptom
A curriculum mechanism intended to grow difficulty adaptively (s1f's ratchet) hit its cap at iteration 248 and never bit again - for 96% of the run its effect was equivalent to constant DR, i.e. the curriculum existed in name only.
Context
When external advice suggested graded wz bands (start ±0.15, then ±0.30), the team agreed with grading but explicitly rejected automatic curriculum, citing the s1f episode. The same logic had already been paid for with push grading: ±0.6 failed twice, ±0.3 was feasible - grading matters, but the grade transitions were made by hand at verified checkpoints.
Change
Ladder policy: difficulty staged manually, one band per rung, each transition gated by the acceptance battery; automatic ratchets not used unless their engagement is monitored and demonstrated.
Outcome
Every C-ladder band change (wz ±0.12-0.25 first, wider later) was an explicit, attributable rung; no silent constant-DR-in-disguise runs recurred.
Mechanism
Adaptive curricula couple their own state machine to noisy training metrics; a ratchet that saturates early stops adapting but keeps its name, so the operator believes difficulty is progressing when it is frozen. Manual staging costs more decisions but each decision is observable and reversible.
Applies when
- choosing between auto-curriculum and staged bands for a new skill
- a curriculum's difficulty parameter plateaus early in training
- post-hoc attribution of what difficulty a lineage actually saw
“C2 的 wz 分级（±0.15 → ±0.30，不要一上来 ±0.6）—— 与我们付过学费的 push 分级同形（±0.6 两轮 FAIL，±0.3 才可行）。但必须手动分级，不做自动课程 —— s1f 的课程化棘轮 iter 248 封顶未咬合，96% 时长等价常量 DR”
train/C_LADDER_RUN.md § 1. 采纳 3 条 (C2 的 wz 分级) A frame-history observation under zero DR memorizes the trainer's plant fingerprint - the estimator must see variation to learn estimation
history-obs-needs-plant-variationIf the observation carries history (stacked frames, RNN), keep at least minimal plant variation (gain/latency jitter) on from the first iteration - "nominal first, robust later" is structurally invalid for estimator-bearing contracts.
Symptom
omni_s1 (fresh 215-dim contract with a 5-frame history window, trained with DR fully off): training all green, yet the MuJoCo gate scored 0/20 on all eight doors - falls within 2 s, seven checkpoints, not one transferred.
Context
The history window exists precisely to let the actor implicitly estimate line velocity and actuator dynamics (the actor is denied base_lin_vel by observation honesty). Under a constant plant that implicit estimator has nothing to estimate - it learns the trainer's exact response fingerprint instead, and any other simulator's micro-differences are out-of-distribution: "5 帧窗按设计就是隐式估计器, plant 恒定时它学到 Isaac 精确响应的指纹". The planned "nominal-first-robust-later" staging was declared STRUCTURALLY incompatible with history observations: "估计器要见过变化才学估计, 否则学背诵" (an estimator must see variation to learn estimation, otherwise it learns recitation). Honest confound note kept: this is mixed with "zero DR does not transfer, period" - but both attributions prescribe the same fix, so no control was run.
Change
S1.1: minimum actuator jitter turned on from day one - kp/kd +/-10%, latency 0-1 frame (friction/COM/mass still nominal, no push - those stay for the S2 ladder).
Outcome
Transfer restored: survival 0/20 -> 20/20, speed 19/20, foot distance 20/20 (remaining failures moved to gait quality, a different disease); the staging doctrine was amended - history-carrying contracts never train under a frozen plant.
Mechanism
A recurrent/history channel fits whatever temporal structure minimizes loss; with a deterministic plant the cheapest structure is the plant's own impulse-response signature, yielding features that are simulator-specific rather than physics-general. Plant variation forces the channel to carry state-estimation features that transfer.
Conflicts
Attribution is explicitly confounded with the simpler "zero DR never transfers" reading ("与「零 DR 本身就不迁移」混杂 … 两种归因处方相同, 不做对照") - the source chose not to spend a control run separating them.
Applies when
- adding frame stacking or recurrence to an actor observation
- a nominal-plant policy fails a cross-simulator gate within seconds
- planning DR staging for a new contract
“frame_hist × 零 DR = plant 指纹过拟合——5 帧窗按设计就是隐式估计器, plant 恒定时它学到 Isaac 精确响应的指纹, MuJoCo 的微小差异即 OOD, 2 s 内摔, 七个 checkpoint 无一迁移。「先标称后鲁棒」的分段与历史观测结构性冲突:估计器要见过变化才学估计,否则学背诵。”
train/OMNI_V0_SPEC.md § 3. S1.1 修订记录 ① A curriculum ramp keyed to the process step counter re-fires on every resume - and shipped policies never saw the penalty
curriculum-counter-lineage-stepsKey every curriculum/ramp schedule to lineage-cumulative progress, not per-process counters; and audit where your shipped checkpoints sit relative to every ramp - a penalty that no product ever experienced is not part of your training.
Symptom
vx+0.30 died at a fixed relative time in every resumed run: resume at 500 -> slide at 1100, zero at 1300; resume at 700 -> slide at 1300, zero at 1500 - absolute depths offset by exactly the resume offset, relative timetable identical.
Context
ramp_reward_weight (the saturation penalty ramp) read env.common_step_counter, which restarts at 0 for every run including --resume. So start_step=600 meant "600 iters after THIS resume", not "lineage iteration 600". The A/B arm pair was the clean proof: their env.yaml differed only in log_dir, only resume point distinguished them, and the omni CurriculumManager had exactly one active term - nothing else could produce that timetable. Second consequence: every shipped checkpoint (s1e-500 at +500, C2-700 at +200, A800 at +100) was selected before its run's +600, so the saturation penalty weight was 0.000 for every product ever shipped - explaining saturation 33% and raw |action| 1.9 against clip 1.0 (hip_roll in bang-bang), i.e. half the heat budget.
Change
Two independent recommendations recorded: (1) make the ramp count lineage steps (add the checkpoint's iteration offset on resume) or pin terminal weights in downstream rungs instead of ramping; (2) give the saturation penalty its own rung - never mixed into a skill-learning rung (that would be two variables again).
Outcome
Explained the recurring +600 death of the highest-amplitude command and the persistent actuator saturation of all shipped products with one root cause; honest caveat booked (at +800 the weight is only -0.086, small, but vx+0.30 is the command demanding the largest action amplitude, so it is squeezed first).
Mechanism
Resumable training splits "the lineage" from "the process"; any schedule keyed to process-local counters silently re-applies its transient to every descendant run, and any product-selection habit that picks checkpoints early systematically samples the pre-ramp regime - the curriculum exists in the config but never in any shipped policy.
Applies when
- resumed/forked training with any scheduled reward or DR ramp
- a metric dies at a fixed offset after each resume
- shipped policies show behavior a late-schedule penalty should prevent
“ramp_reward_weight 读的是 env.common_step_counter，它每个 run 从 0 开始，--resume 也不例外。… 原始 run / 臂B | 500 | 1100 = +600 | 1300 = +800；臂A | 700 | 1300 = +600 | 1500 = +800 … 所有出品其实从没见过饱和罚。… 这解释了 sat_max_pct 33%、raw |a| 最大 1.9（clip 是 1.0）—— hip_roll 一直在 bang-bang，而罚它的那一项权重恒 0。热账的一半在这里。”
train/C_LADDER_RUN.md § 3g. 系统性问题：saturation_ramp 每次 resume 归零 The deploy-side walk/recovery switch - into recovery at tilt > 65 deg held 0.3 s, back at tilt < 15 deg with angular rate < 1 rad/s and straight knees held 1 s, a 15 s timeout, last action cleared both ways - and the handoff steps the design required are only partly implemented
walk-recovery-fsm-handoffSpecify a deploy-time controller switch as hysteretic, time-filtered predicates the robot can measure (proxy what it cannot, e.g. straight knees for height), a timeout that ends in a safe stop, and a complete handoff (history, clock, last action, command ramp) - then test that the code performs every handoff step, because the design document is not the implementation.
Symptom
With a recovery policy and a locomotion policy as separate networks, the robot needs a switch: when is it "fallen", when is it "up", and what state must be reset so the next policy does not act on the previous one's history.
Context
The 08-09 design: enter recovery when fallen (tilt > 55 deg or height < 0.60 x 0.384 m) for 150 ms, leave for a stand-hold when upright (tilt < 12 deg, height > 0.85 x 0.384 m, feet steady, |omega| < 0.8) for 400 ms - wide entry, strict exit, hysteresis - then a mandatory handoff trio before walking resumes (reset the walking policy's observation history, restart its phase clock at 0, clear its previous action and latency buffer) and a command ramp instead of a jump. The 08-14 implementation in deploy_policy (--recovery-policy): each policy under its own manifest contract (walk: nominal + scale; recovery: beta-anchored), the same rl_default gains, the tilt cutoff disabled; RECOVERY at tilt > 65 deg for 0.3 s; LOCO again at tilt < 15 deg and |omega| < 1 and knees straight (< 0.35 rad) for 1 s - the robot's computer has no height estimate, and straight knees stand in for height so a V3.0-style upright kneel cannot pass as standing; RECOVERY longer than 15 s ends in a safe stop; last_action cleared on both switches; command forced to 0 during RECOVERY; power scaling applies to LOCO only.
Change
An open account was written down with it: the recovery end state (0.355 m stance, hip yaw -/+27 deg) is outside the walking policy's training start distribution, so the first test must use stand / zero command as LOCO, and the long-term fix is to widen the walking policy's initial states rather than bend recovery's stance to suit walking.
Outcome
The spec records only a successful compile (py_compile); the hang test was left to be done on site. The operator runbook carries the three-step procedure (hang with stand as LOCO, mat and push, then omni walk as LOCO) but no outcome.
Mechanism
Two policies trained separately each assume their own history, clock and last action; a switch that carries any of them across feeds the next policy a state it never saw - the design called this "the walking policy seeing a ghost history".
Conflicts
The 08-09 design requires resetting history, clock and previous action plus a command ramp; the 08-14 implementation records last_action clearing and a zero command during recovery; the one-leg spec of 2026-09-14 lists the handoff hygiene as specified but not implemented - reset_history() is called by nothing (a 215-dim policy would carry four frames of pre-fall history), the phase clock is not zeroed, and there is no command ramp back to LOCO. No hardware run of the FSM is recorded in either source.
Applies when
- switching between separately trained policies on hardware
- a policy with history or phase observations is re-enabled mid-run
- the robot lacks a sensor the switching criterion was designed around
“**判据**:进 RECOVERY = 倾角 >65°(`--fall-tilt-deg`)持续 0.3 s;回 LOCO = §5 真机可测子集:倾角 <15° ∧ |ω|<1 ∧ **膝直 <0.35 rad(NX 无高度观测, 高度门用膝直代理 —— 防 V3.0 型"跪坐但直立"误判)** 持续 1 s (`--recover-hold`);RECOVERY 单次 >15 s(`--recovery-max-time`)安全停。 … **切换卫生**:两向切换 last_action 清零;RECOVERY 态 cmd 强制 0”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §50 FSM 双策略调度(2026-08-14,用户令):deploy_policy --recovery-policy A pull-assist curriculum keyed to a global success share was satisfied by the categories that already worked and withdrew before prone learned anything - conditioning the criterion on prone took it from 2.5% to 98.7%
curriculum-criterion-conditioned-on-lagging-categoryMeasure a curriculum's advancement criterion on the population the scaffold is meant to help; a global success share is met by whatever already works, and the help is withdrawn before the lagging case learns.
Symptom
Under the beta-anchored action, supine and side stood reliably while prone still sat (1.3%). A pull-assist curriculum added to help it was withdrawn completely within ~790 iterations and prone moved only to 2.5% (noise).
Context
The pull assist follows HoST: an upward force on the base, active only when the torso is within 30 deg of vertical, scaled by body weight (HoST's 200 N on G1 = 0.583 BW -> 56 N here, steps of 5.6 N, ten levels to zero); the product must pass with no assist. It had already taught sit-to-stand in V1. In V2.1 its advancement criterion was the standing-time share over all envs (threshold raised to 0.55 because the share was already ~0.53).
Change
V2.2: PullAssistForce with gate_category = "prone" - only envs whose first step classifies them as prone count toward the criterion - and the threshold back at 0.35. A feasibility signal was pre-registered: if prone's share stayed near zero under the full 56 N, return to the roll-over path instead of adding force.
Outcome
The prone-conditioned curriculum withdrew level by level only as prone itself passed: prone 98.7%, and the four-category gate passed for the first time on the line (98.6% overall, re-falls 0%, torque gate PASS).
Mechanism
A pooled success share is filled by the categories that already succeed (supine/side ~53%), so the scaffold is removed on their account before the lagging category has used it.
Applies when
- an assist, guide force or easier setting is withdrawn by a success threshold
- one task category lags while the pooled metric looks healthy
- a curriculum ran to completion without changing the lagging category
“**教训:全局站立占比阈会被存量类别(supine/side ~53%)凑够,拉力在 prone 学会前就撤光了 —— metric 设计失误,不是拉力机制失效**(它在 V1 教会过 坐→站)。 … **V2.2(已启动)**:`PullAssistForce` 加 `gate_category="prone"` —— 达标判据 只统计 prone 类 env”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §34 V2.1 判决(2026-08-10) Anchoring the action on the measured joint angle (target = q + beta*a), with a beta curriculum down to tau_limit/kp, bounded torque by construction, removed the re-falls and later stood the robot up on hardware
beta-anchored-action-targetFor large-motion skills on position-controlled actuators, bound the action relative to the measured joint angle with a per-joint authority of tau_limit/kp, curriculum the authority down from full range, keep the curriculum state out of the observation and pin acceptance at the deployed authority - and make the deployment code refuse to run the anchored contract without a measured q.
Symptom
The V0 full-range absolute action produced violent targets; the V1 command-anchored rate limit made standing oscillate. Both failure modes came from how the action becomes a target.
Context
V2.0 (user approved, from scratch): BetaAnchorJointPositionAction, target = q_measured + beta_j(m)*a, memoryless per step. beta_j(m) = floor + m*(beta0 - floor), beta0 = the contract half-range (m = 1 reproduces V0 authority), floor = min(tau_limit/kp, beta0): hip_pitch 1.309 -> 0.40, knee 1.047 -> 0.40, hip_yaw -> 0.917, the other joints unchanged - the tightening lands exactly on the joints the V0 torque account convicted. m drops 0.1 per step when a standing-share EMA exceeds 0.35. beta is NOT in the observation, so the 45-dim contract is untouched; acceptance is pinned at m = 0 because the Python curriculum state is not saved in the checkpoint. The deployment chain got a new profile (recovery_v2: action_anchor current_q, explicit per-joint beta written into the contract, independent of the gain profile), and policy_io raises if q is missing rather than silently falling back to the absolute contract; the old profile's check reproduced its pre-change deviation bit for bit.
Change
New action term and beta curriculum; later the RS06 floor was lowered 0.40 -> 0.30 -> 0.25 (kp*beta 7.5 N*m) and the stamped deployment profile was synced to 0.25.
Outcome
First acceptance at m = 0 (v2_0b): re-falls 0% in every category, the torque gate passed for the first time on the line (worst 69.9%), knee jitter 0.004; supine 98.8 / side 88.8% with prone and mid still failing (fixed by the conditional pull curriculum). MuJoCo showed demand at or under the limits (hip_pitch 11.7/12 against V0's 26.8). Lowering beta cut impact (hip_pitch demand 9.7 -> 8.5 N*m) but barely slowed the get-up - it had become coordination-limited. Enabling the policy moves the target only +/-beta around the current pose, so there is no homing fling; the 08-11 real get-up and the later v3_1p1c both run on this contract.
Mechanism
kp*beta caps the proportional torque in a single step with no build-up delay and no memory, giving both a hard impact bound and full balance bandwidth.
Applies when
- a skill needs full joint range but hardware torque limits are low
- absolute position targets cause impacts or saturation
- changing the action semantics of a contract that deployed policies share
“**动作项** `BetaAnchorJointPositionAction`:`target = q_实测 + β_j(m)·a`, 逐步无记忆 … **Play/验收钉 m=0(= floor = 部署档)**:python 课程状态不进 checkpoint, Play cfg 显式 `beta_m_start=0` … 判读:**结构赌注兑现** —— 站姿零再摔 + 力矩账首过(kp·β 封顶按构造)”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §33 V2.0 预注册(2026-08-10,用户点头开工):β 锚定动作空间,从零训 The best checkpoint to SHIP is not the best checkpoint to CONTINUE FROM - maturity is capital against adaptation shock
root-maturity-vs-product-qualityDecide shipping points and fork roots separately: gates rank products, but a root candidate must prove itself by surviving a continuation under the next rung's shift (dual-arm if in doubt) - and prefer the more-trained point as root when product metrics conflict with maturity.
Symptom
A band re-audit found s1e-300 beat the incumbent root s1e-500 on nearly every quality gate (stepping 19/20 vs 13/20 with historically-best 26.9 mm swing, speed gate 14/20 vs 2/20, heading 26 vs 54 deg/20 s) - suggesting the root had been mis-picked and the younger point should take over.
Context
The dual-arm control settled it the other way: continuing the S2 PD rung from s1e-500 adapted smoothly (3/3 smoke throughout), while the b300 control arm (same config, from s1e-300) fell into a survival valley under the PD shock (+100 iters: 1/3 -> 0/3), never climbed out within budget, and its 800-iter product scored 13/20 survival - eliminated. Verdict: "幼年点自身指标再好也扛不住新 DR 适应冲击, 成熟度是本钱,s1e-500 根被数据背书" - a young point's own metrics, however good, do not survive new-DR adaptation shock; maturity is capital. The audit still yielded value: the band scan (200-1000, per-100) mapped the lineage's arc (200 dragging -> 300 peak -> 400+ decay -> 900+ drift blowout), and 300 remains the better PRODUCT answer for shipping-as-is questions.
Change
Selection doctrine split into two questions with different answers: best-product point (quality gates at the point itself) vs best-root point (survives adaptation shocks; more training age = more capital), each decided by its own evidence - and root claims settled by a dual-arm continuation test, not by point metrics.
Outcome
s1e-500 kept the root role with data behind it; the S2e ladder built on it passed rung after rung, while the b300 line was closed at the cost of one control arm.
Mechanism
Early checkpoints sit near sharp optima with less accumulated robustness structure; their headline metrics reflect the narrow training distribution, not resilience to distribution shifts. A continuation rung is itself a distribution shift, so the root property being selected for is shock tolerance - observable only by actually continuing, never by static gates.
Applies when
- a younger checkpoint outscores the current root on quality gates
- choosing the base for a robustification or command ladder
- a continuation run stalls in an early survival valley
“b300 对照臂 … PD 冲击下存活谷(+100 起 1/3→0/3),预算尽未爬出,800 档 20-seed 存活 13/20 出局——幼年点自身指标再好也扛不住新 DR 适应冲击,成熟度是本钱,s1e-500 根被数据背书”
train/README.md § omni_s2e_pd (b300 对照臂) / s1e 选点重审 Curriculum-gate a penalty to the phase where its disease occurs - early on it only taxes exploration
gate-penalties-to-the-disease-phaseFor penalties aimed at late-stage pathologies (freezing, saturation, degenerate attractors), ramp the weight in only after exploration noise has decayed; anchor the terminal weight to measured healthy-vs-sick raw values, and shift all related tripwires to after the ramp completes.
Symptom
The action_saturation penalty, applied from iteration 0 in v8a, taxed exploration itself: with init_noise_std 1.2 the sampled actions paid ~-2.45/step before any policy had formed - while the disease it targets (clamp freezing) is a LATE pathology (v9 froze at iteration ~2624).
Context
v10 re-introduced the same penalty behind a curriculum gate: weight 0 until iter 1000, ramping linearly to -1.0 by iter 2000 - present only when the disease can occur, absent while exploration noise dominates. The trust argument was evidence, not hope: in v8a the term, while active, had pulled joint_pos_ref from 0.041 up to 0.155 and climbing - proof it can extract a policy from the frozen pit. Weight magnitudes were anchored to measured raw values (healthy v5 0.310 / v6 0.106 vs frozen v7 1.145 / v9 1.22 per step: at -1.0 healthy pays 6-18% of tracking, frozen pays 60%+, standing ~0). v10c then isolated the gated term as THE anti-freeze mechanism by single variable, upgraded to untouchable status in v11: "S 的门控机制(v10c 单变量铁案:任何情况下 不许撤,只许调终值)" - and v11 dared to relax other penalties only because S stood guard.
Change
action_saturation gated 0 -> -1.0 over iters 1000-2000 (later terminal value tuned -1.0 -> -0.5 with the gate mechanism itself frozen); tripwires adjusted to respect the gate's timing (freeze check moved to iter 2500-3000 to give the ramped term its effect window).
Outcome
Freezing stopped recurring while early training kept full exploration; the mechanism graduated from experiment to invariant within two versions.
Mechanism
A penalty's incidence depends on who occupies its support: early in training that is exploration noise (whose suppression starves learning), late it is the converged pathology. Time-gating aligns the penalty's presence with its target's presence, buying the constraint without the exploration tax - and tripwire timing must then be computed from the gate schedule, not from ungated precedents.
Applies when
- a structural penalty punishes exploration in early training
- a late-onset pathology (freeze/saturation) needs a standing guard
- deciding when a curriculum ramp should engage
“v8a 实锤它的病根是"罚在采样动作上"——init_noise_std 1.2 的早期等于罚探索(~−2.45/步);而冻结是晚期病(v9 速率 2624 才死平)… 门控让它只在病发期在场。… v8a 里它在场时 joint_pos_ref 从 0.041 爬到 0.155 且仍在升——有从低谷爬出的实证。”
train/WALK_V10_SPEC.md § 2. S 保险 —— action_saturation 课程门控 Design the next run to complete the 2x2 - either outcome then convicts or acquits a factor cleanly
fill-the-missing-factorial-cellWhen two config factors are jointly suspected, lay out the factorial of existing evidence, spend one run on the missing cell with both interpretations and an early-abort tripwire written in advance - and treat either outcome as a verdict, not a disappointment.
Symptom
Hip joints froze at the action clamp in v7, but the history could not say whether the culprit was the raised action_rate (-0.2) or the halved reference amplitude (scale 0.15): existing versions covered only three corners of the (rate x scale) space - v5 (-0.03, 0.30) healthy, v6 (-0.03, 0.15) healthy, v7 (-0.2, 0.15) frozen.
Context
v9 was designed explicitly as the missing cell (-0.2, 0.30), with the readings pre-registered: v9 not frozen -> the real anti-freeze force was always the reference amplitude and -0.2 may stay; v9 frozen -> -0.2 is convicted beyond appeal (freezes at both amplitudes) and the next version goes straight to a structural fix. "两个结局都是干净的信息" - both endings are clean information.
Change
One training run allocated purely to complete the factorial, with freeze tripwires (joint_pos_ref telemetry <0.1 at iter 1000-1500 -> abort, do not run to 6000) so a conviction costs the minimum compute.
Outcome
v9 froze - the rate weight was convicted at both amplitudes ("−0.2 铁案定罪"), and v10 moved to the structural saturation fix with the weight question closed instead of re-litigated.
Mechanism
Three corners of a 2x2 leave the two factors confounded in the failure corner; the fourth observation makes each factor's marginal effect identifiable. Pre-registering both readings turns the run into a guaranteed-informative experiment regardless of outcome.
Applies when
- two config changes are confounded in a failure
- version history already covers some corners of a factor grid
- deciding what single experiment buys the most attribution
“这恰好补齐一个 2×2 实验矩阵的缺格 … v9 不冻 → 真正的抗冻结主力一直是参考摆幅,−0.2 可以留;v9 仍冻 → −0.2 铁案定罪(两种摆幅下都冻),v10 直接上结构修复 … 两个结局都是干净的信息。”
train/WALK_V9_SPEC.md § 0. 设计原则 (2×2 实验矩阵) The get-up kept getting faster because standing earlier paid more every step - lowering torque authority barely slowed it, and only zeroing the standing income for the first 3 s moved the pace into the design band
per-step-income-drives-speed-time-gateWhen a skill is too fast, find the term that pays for finishing early and gate that income by time; keep the "get into position" term ungated so the policy does not learn to wait, use a ramp instead of a cliff, and confirm with a paired same-level experiment that the drift is motivational before changing it.
Symptom
The user judged the get-up too fast (Isaac medians about 0.7-1.6 s) and suspected path dependence: the policy seemed to get faster the longer it trained.
Context
Lowering the beta authority 0.40 -> 0.30 cut impact but moved supine only 1.70 -> 2.00 s: coordination-limited, not torque-limited. A paired experiment inside one beta level (checkpoint 15,600 vs 18,499, +2,900 iterations, same ruler) measured the drift: get-up medians -7 to -10%. The spec concluded the motive, not the path, was the cause - per-step standing income pays for every early step, and any lineage (even one from scratch) races toward the fastest solution inside its constraints.
Change
V2.5: the standing income (base_height, stand_pose, still, feet_on_ground) multiplied by w(t) = clamp(t/3 s, 0, 1); upright deliberately NOT gated, so righting and sitting up early still pay and the policy is not taught to lie flat and wait; a ramp, not a step. V2.5b: zero before t0 = 3 s, then a 1 s ramp.
Outcome
V2.5: Isaac 100%, get-up +17-43% slower, MuJoCo 100/100/98/100% (the best cross-simulator reading yet), still short of the 3.5-4.5 s design band - a linear ramp only discounts early income. V2.5b: MuJoCo supine 2.04 -> 4.10 s and prone 3.18 -> 4.04 s, inside the band; the Isaac pace barely moved (a lineage habit on a gradient-free plateau). Later the zero gate proved harmful when trained from scratch (curriculum-history-is-part-of-the-product) and in the V3.1 lineage (time-gate-vs-wide-stance-retire-the-fix).
Mechanism
Constraints on authority or velocity change how the fastest solution looks; the time structure of the task income decides how fast the fastest solution is.
Applies when
- a policy is faster or more aggressive than wanted and constraints do not slow it
- progress-style rewards pay every step spent at the goal
- performance drifts faster with more training at fixed settings
“**V2.5 机制(唯一)**:站立收入(base_height/stand_pose/still/feet_on_ground) 乘时间斜坡 w(t)=clamp(t/T_gate,0,1),T_gate=3.0 s;**upright 刻意不门控** (翻正/坐直早期照常拿钱,防"躺平等门开" … **用户假设量化 证实:逐步计酬动机在 β 包络内持续压缩时间,约束挡不住动机 —— V2.5 动机层 修法为正解。**”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §39 V2.5 预注册 / §39 补 配对实验 PPO's Gaussian noise cannot compose phase-locked oscillations - deliver them as feed-forward and let the policy learn the residual
feedforward-for-phase-locked-skillsIf a skill needs a temporally coherent (phase-locked) action component, do not expect step-wise exploration to find it: inject a verified feed-forward and train the policy as a residual stabilizer, keeping the feed-forward inside the deployment contract.
Symptom
Four different reward arrangements (no reference / wrong-sign reference / correct-sign reference / cage released) all failed to elicit sidewalk, while open-loop probes proved the behavior existed and was safe on the same platform with the same policy as base.
Context
Producing lateral velocity requires a phase-locked hip_roll oscillation synchronized to the gait clock. PPO's exploration is per-step, zero-mean, uncorrelated Gaussian noise - it can never compose a sustained phase-locked component, so the behavior is unreachable by exploration regardless of how it is rewarded. The fix changed the delivery channel: target = default + scale*action + lat_ff(cmd_vy, phi). The policy's action becomes a residual on top of the feed-forward, retaining full balance authority (it can even cancel the feed-forward); the feed-forward supplies exactly the component exploration cannot. This mirrors why the sagittal joint_pos_ref worked (it also delivered phase structure), just via a different channel.
Change
Contract-level change, done cleanly: new profile omni_ff (= omni + lat_ff_gain -0.5), existing omni profile bit-identical; feed-forward applied after the action delay stage; missing cmd/phase raises instead of silently dropping; deployment must use the same phi as build_obs (recomputing gives a one-tick phase misalignment).
Outcome
From C2-700, +100 iterations sufficed: product omni_c4_ff800 scored vy +120%/+125% (from +4%/-1%), 260/260 cells at 20/20 survival, zero old-skill regression, left/right gap 5 pp - the entire C4 saga resolved by changing the delivery mechanism, not the reward.
Mechanism
Exploration noise spans only the subspace its correlation structure can express; skills requiring coherent oscillation lie outside the span of i.i.d. per-step noise. Feed-forward moves the required structure into the action pipeline where it needs zero probability mass to appear, reducing the learning problem to stabilizing around a demonstrated behavior - which PPO does well.
Applies when
- a periodic/oscillatory skill trains flat under every reward variant
- open-loop injection of the behavior already works
- considering GRU/curriculum/exploration tricks for a rhythmic skill
“病因不在奖励，在探索形式：产生侧向速度需要相位锁定的 hip_roll 振荡，PPO 的逐步高斯噪声零均值无相关，合不出相位锁定分量。… target = default + scale·a + lat_ff(cmd_vy, φ)。策略动作因此是前馈之上的残差，保留全部平衡权限”
train/C_LADDER_RUN.md § 3j. C4-redo4：唯一变量 = 侧步参考改为前馈注入（契约级） A policy's gain profile is part of its contract - the one-leg policy needs per-joint gains the default profile lacks, and the manifest refused an evaluation under the default once; the recovery contract's beta was never stamped, a known gap not to repeat
gain-profile-belongs-in-the-stampStamp everything that defines the closed loop a policy was trained in - gains included - into its manifest, and make every consumer refuse a mismatch; a profile field that is not in the stamp is a silent misconfiguration waiting for an operator to forget a flag.
Symptom
A policy trained with hip_roll kp 80 and ankle_roll kp 60 behaves differently, or falls, under the default kp 20/12 profile - and the gain profile is a command-line flag an operator can forget.
Context
The one-leg line added a gain_profile field to the contract so the stamped manifest carries it; the spec's deployment note says the manifest guard blocks rl_default and that it had already bitten once in simulation (an evaluation run without the one-leg profile). The same spec states the general rule - any new profile field must be synced into the manifest builder - and names the counter-example: the recovery line's beta was never put into the manifest. The recovery line itself had decided that its anchored authority is computed from the base rl gains and written into the contract so it cannot drift with the gain flag, and that the older kp x 0.9 profile chosen in the V0 era does not match the beta contract and must not be used.
Change
Gain profile as a contract field checked at load; per-contract gain choices written into the run sheets.
Outcome
Evaluations and hardware runs of the one-leg policy run under rl_oneleg or are refused; the recovery beta gap stayed recorded as known.
Mechanism
A policy is trained against a closed loop whose gains are part of the plant; running it under other gains is an out-of-distribution plant, exactly like a wrong observation scale.
Applies when
- a skill introduces per-joint or skill-specific gains
- deployment gains are chosen by a command-line flag
- adding any new field to a policy profile
“增益档 `--profile rl_oneleg` 必须给 —— manifest 防线会拦 `rl_default`（sim 已咬合一次） … （recovery 的 β 未进 manifest 是已知缺口，不再复制）”
git:Lucen V2@origin/oneleg-line:train/ONELEG_V0_SPEC.md § §9b AGX 真机手顺 要点 / §3 契约 Drop the frozen policy into chosen configurations - a squat 2.7 cm lower than the stuck pose stood 52% of the time, the stuck W-sit 0%, and the interpolation between them showed a wall, not a slope
configuration-probe-wall-not-slopeWhen a policy is stuck, probe the frozen policy from a grid of hand-placed start configurations, including interpolations between the stuck state and a nearby state it escapes from; one read-only experiment separates height, torque, sampling and configuration and tells you whether to prevent entry or train the exit.
Symptom
After R0.3 the policy stood from 62% of starts and never from the W-sit it fell into; height, torque, missing samples and reward were all plausible suspects.
Context
A read-only probe placed the R0.3 policy directly into specified configurations. Squats (hip, knee, ankle) = (-.65,-1.3,-.65) stood 100%, (-1.0,-2.0,-1.0) 89.8%, (-1.2,-2.4,-1.2) at 0.176 m 52.3%; the measured W-sit at 0.203 m 0.0%; the account-(3) hand-over state (146 deg tilt) 34.4%; linear interpolations from the W-sit toward the squat at 25/50/75% stood 0.0/0.0/3.1%. The squat family's quasi-static torque is 16% of the limits, and the W-sit was visited ~9 s per episode in training. FK showed the squat family (-a,-2a,-a) keeps the torso vertical, the feet flat and the COM over the feet all the way from 0.146 m to 0.384 m.
Change
Height, torque and sampling were eliminated in one experiment; the next rungs targeted entering the W-sit (foot placement) instead of escaping it, and seeding the dead point itself was ruled out because it was already visited every episode.
Outcome
Pure configuration: the W-sit (hips externally rotated +/-47 deg, knees folded 110 deg, shins flat, feet beside the body) is a different place from the sagittal squat (feet flat under the COM). The policy's standing skill was bound to a narrow sagittal family, and the wall was confirmed by the interpolation. The foot-placement rungs that followed took prone from 0/159 to 158/159.
Mechanism
A learned skill covers the neighbourhood of the states it succeeded from; a start state outside that neighbourhood fails regardless of height or torque, and an interpolation that stays at zero until close to a working state shows the boundary is sharp.
Applies when
- a policy stalls in a specific posture and several causes are plausible
- deciding between reverse-curriculum seeding and entry-prevention shaping
- a feasibility account says a path exists but the policy does not take it
“**决定性对比:比死点矮 2.7 cm 的蹲姿站立 52.3%,死点 0.0%。** 所以不是高度、 不是力矩(蹲姿族准静态力矩膝 1.96/12、踝 1.24/17,只占 16%)、也不是训练采样 (死点每局被访问 ~9 s)。**是纯位形问题** … 插值实验进一步显示这**不是坡是墙** —— 走到 75% 仍只有 3.1%”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §15 死点位形实验(只读探针,同一个 R0.3 策略放进指定位形) Fall recovery was defined as the whole chain - any fallen pose, a stable stand, a clean hand-back to walking - and built as a second policy behind a deploy-side switch, not folded into the walking PPO
recovery-two-policies-and-a-state-machineDefine a recovery skill by the whole chain it must complete, including the hand-back to the next controller; if it is built as a separate policy, make the switching logic and its handoff contract a deliverable of their own, and keep the recovery observation contract a subset of the locomotion one so a unified policy stays possible later.
Symptom
A walking robot that falls needs a human to stand it back up. The design question on 2026-08-09 was whether to teach getting up inside the existing omni walking policy or beside it.
Context
The user set the goal as "any fallen pose -> stand up alone -> stand stably", and the spec named the real difficulty as the full chain fall -> recovery -> stable stand -> correctly initialised walking history and clock -> walking, making the deploy state machine a first-class deliverable. A unified single policy had a real-robot precedent (arXiv:2605.18611, a state-dependent gate near 37 deg tilt) but was deferred until a recovery policy and an omni policy were each reliable. The line ran on its own branch and worktree with every walk/stand/omni/run config path untouched. The development path copied the G1 learned get-up logic (arXiv:2502.12152): first find any feasible get-up (ugly accepted), then add smoothing, torque and real-robot constraints. The recovery contract kept the base 45-dim observation (command slice held at 0, no gait phase, no frame history - their reasons do not apply to a skill without a clock or a velocity task), so it stays a prefix of the 215-dim omni contract and a later merge is not foreclosed.
Change
Two policies and a deploy-side switch instead of one retrained walking policy; recovery got its own minimal contract (45 dims, full-range action, later the beta-anchored profile) and its own acceptance battery.
Outcome
The split held for the whole line: on 08-14 deploy_policy gained a second (PolicyIO, ONNX) pair behind --recovery-policy, each loaded under its own manifest contract, and the runbook runs stand_v1b or omni_c4_ff800 as the locomotion side with recovery_v3_1p1c. The literature scan of 08-10 found that every verified get-up implementation deploys one end-to-end policy (or softly gated experts) and stages only on the training side - so the runtime state machine here is the walk/recovery switch, not a staged get-up.
Mechanism
A separate policy keeps each reward table single-purpose and lets a proven walking lineage stay byte-frozen; the cost moves to the handoff, where every piece of state one policy leaves behind (history, clock, last action, command) must be reset for the other.
Applies when
- adding fall recovery or get-up to a robot that already walks
- choosing between one unified policy and a switched pair of policies
- designing the observation/action contract of a secondary skill
“先做 recovery policy + omni policy 两个策略,部署侧状态机切换;不把 recovery 硬塞进现有 omni PPO。 … 任务定义:**任意跌倒姿态 → 自己站起来 → 稳定站立**。真正的难点不只是"起身", … omni walk**(§6 部署状态机是本 spec 的一等公民,不是附录)”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §0 目标口径与架构判决(用户 2026-08-09 定) Narrowing the speed range to stop high-speed falls entrenched crouch-shuffling - judge gait quality at the speed that demands a gait
low-speed-commands-reward-draggingSet command ranges to include speeds that physically demand the target behavior, and evaluate behavior-quality gates at those speeds; when a restriction is added to suppress a failure, check what new optimum it creates at the remaining commands.
Symptom
After the command range was narrowed to (0.15, 0.35) m/s (to treat walk_v1's 134% overspeed and 8.3 s fall at 0.5), the policy settled into crouched foot-dragging; tracking rose monotonically with speed (63% at cmd 0.2, 76% at 0.3, 87% at 0.45), showing low speeds were where the degenerate gait was optimal.
Context
The narrowing advice was the author's own and is retracted in the file: it treated the symptom (falls at speed) while reinforcing the root cause (at 0.15-0.35 m/s, shuffling in a crouch is globally optimal - the Froude number is so low that even humans would not lift their feet). A zero-cost experiment confirmed the flip side: at cmd 0.5 the same policy met BOTH tracking (81%) and clearance (23.0/23.2 mm) standards.
Change
Speed range widened back toward (0.15, 0.5) - upper bound deliberately slightly above the mechanically feasible ~0.44 m/s so the policy finds the boundary itself; acceptance re-pointed: gait-quality criteria (tracking, clearance) judged at 0.45-0.5 m/s, low speed kept only as a survival check.
Outcome
v4 -> v6 progression under the widened range delivered 87% tracking with 34 mm clearance; the "low command = drag" account was confirmed by the monotone tracking-vs-speed curve.
Mechanism
Command distribution is part of the reward: physics prices gaits per speed, and at very low speed the energetic optimum is no swing phase at all. Restricting training to that regime makes the degenerate gait the correct answer to the posed problem - and grading a gait at a speed that does not require stepping measures nothing.
Applies when
- a gait degenerates after a command-range restriction
- quality metrics improve monotonically toward the range boundary
- writing acceptance criteria for gait quality vs survival
“现在看那个建议可能起了反作用:0.15~0.35 m/s 下蹲着蹭就是全局最优,抬腿反而亏。收窄治的是"高速摔倒"的症状,却强化了拖地的病根。… 验收标准里的 cmd 0.2 本身就是拖地速度(Froude 数极低,人在那个速度下也不抬脚)。accept_v2 应把速度跟踪与 clearance 的判定点改到 0.45~0.5 m/s”
train/WALK_DIAGNOSIS.md § ② 放宽速度区间 / ① 零成本实验 Under continuous 3-axis uniform sampling, pure straight-line walking is a zero-measure event the policy never trained
zero-measure-commands-need-mode-samplingEnumerate the exact command points users will actually issue (straight, stop, in-place turn) and give each explicit probability mass via mode sampling with off-axes pinned to zero - never assume a continuous sampler covers its measure-zero subsets.
Symptom
"The robot drifts even in sim when told to walk straight" persisted across reward tunings - because with commands drawn as vx in [0.15,0.5] x vy ~ U(+/-0.2) x wz ~ U(+/-0.6), the event vy=0 AND wz=0 has probability zero: pure straight-line walking was never sampled even once.
Context
Restart evidence item #3: "纯直行是零测度点 … 'sim 里直行就漂'是分布的 必然,不是 reward 没调好" - the drift metric was legitimately drowned by commanded turning (v11's own comment self-documented this). The structural fix is discrete mode sampling: a custom ModeVelocityCommand that first draws a mode by share (stand/forward/back/turn/side/mixed), then draws values only on that mode's axes with all others pinned to exact zero - which is also what preserves single-variable discipline in the C ladder (native 3-axis uniform "采不出'离散模式桶' … 把 C1~C4 的单变量纪律直接毁掉"). The mixed mode later got an ellipsoid constraint rather than a cube for the same reason in reverse - corner combinations of a cube are unrepresentative extremes.
Change
Command generation moved from independent per-axis uniforms to mode-bucket sampling with pinned-zero off-axes (plus 20% rel_standing); acceptance likewise evaluates per mode.
Outcome
Straight-line behavior became a trained, testable mode instead of a measure-zero hope; the C ladder could add one mode per rung with provable isolation.
Mechanism
A policy optimizes expected reward under the command distribution; events of probability zero contribute nothing to the objective, so exact-zero-command behaviors (straight walk, stand, in-place turn) are only learned if the sampler gives them mass. Product-of-uniforms distributions concentrate mass on mixtures and give none to the pure behaviors users actually command.
Applies when
- a "simple" command (straight, stop) underperforms mixtures in sim
- designing command distributions for velocity-tracking tasks
- a ladder needs per-mode isolation for attribution
“纯直行是零测度点:最终 command 为 vx∈[0.15,0.5] × vy∈U(±0.2) × wz∈U(±0.6) 连续均匀,vy=0∧wz=0 从未被专门采样 —— "sim 里直行就漂"是分布的必然,不是 reward 没调好 … Isaac 原生 UniformVelocityCommand 是三轴各自 uniform,采不出"离散模式桶"”
train/OMNI_V0_SPEC.md § 0. 为什么从零 (3) / 三件前置 (1) Sort external training advice into adopt / already-have / modify / would-trap by recomputing it on your own config
external-advice-audit-against-own-arithmeticNever apply external tuning advice directly: recompute each claim on your own reward table and probe data, classify it adopt / have / modify / trap, and record why - and verify external citations actually exist.
Symptom
External AI/literature advice for the omni ladder arrived plausible-sounding but was written without knowledge of this robot's actual reward table, contract, and history; following it blindly would have broken single-variable discipline and, in one case, made sidewalk unlearnable.
Context
Before the C ladder, every external suggestion was audited: 3 adopted (ellipsoid command sampling; staged wz bands; command-switch acceptance), 3 already present (unified reward; frame history - the frozen 168-dim 5-frame window; per-100-iter acceptance), 2 modified (stand share kept at 20% to avoid a second variable; back share NOT raised because the probe showed backward works untrained 20/20@67%, so oversampling would only crowd out forward), and 1 flagged as a trap: "start vy very small (0.06-0.15)" - on THIS reward table vy was only an L2 tax, so ignoring a vy=0.06 command costs 0.4% of the vx tracking scale, 28-180x cheaper than ignoring forward, with quadratic shrinkage making small commands weaker still. A separate retrieval-reliability note: two search agents returned fabricated verbatim quotes from arXiv PDFs (2 papers, verified fake and discarded); only HTML/abstract/source-verifiable material was used.
Change
Advice classified only after recomputing each claim with local numbers; the "start small" trap was replaced by adding a gated lateral tracking term (the ladder's only true reward surgery) instead of shrinking the command.
Outcome
The adopted items (ellipsoid modes, staged wz, transition acceptance) entered the ladder; the trap was avoided; one external factual error (calling s1g the mainline start - it was falsified 0/20) was caught. Later, one initially-dismissed item (sigma=0.15 too narrow) turned out right for a different reason than claimed - see cycle-average-tracking-for-gait-quantities.
Mechanism
External advice encodes the advisor's reward table and robot, not yours; the transfer-validity test is whether the claim survives recomputation under your own arithmetic (reward margins, probe baselines, contract freeze). Items that survive become experiments; items that don't become documented traps.
Applies when
- incorporating LLM or literature advice into a training plan
- advice conflicts with locally measured baselines
- an external claim depends on reward-table details the advisor cannot know
“其建议 C4「先很小，vy = ±0.06~0.15」—— 在我们这张奖励表下会让侧走学不起来 … 忽略侧走比忽略前进便宜 28~180 倍，且指令越小激励越弱（平方缩放）—— "先很小"在稳定性上对、在梯度上正好把信号缩没了。”
train/C_LADDER_RUN.md § 1. 外部 AI 训练建议的评估（采纳 / 已有 / 要改 / 会踩坑） Freeze the deployment contract, stamp every export, and let an automated checker catch wiring bugs
contract-freeze-and-checkerFreeze and fingerprint the policy I/O contract; ship contract changes as new versioned profiles that leave old artifacts bit-identical; and extend the automated contract checker with every pipeline change, forcing the new path to execute in the check.
Symptom
Contract-level changes (observation layout, action pipeline) are where silent sim/real divergence is born; two real wiring bugs appeared the one time the action pipeline was extended.
Context
The 215-dim observation contract was frozen ("纪元 3，三机 digest" - an era number plus a digest agreed across three machines); proposals that would break it (e.g. a GRU memory) were rejected on contract grounds. Every exported ONNX is stamped and verified with a manifest (onnx_manifest --stamp / --verify), and deployment refuses mismatched combinations. When C4 added the lateral feed-forward, it went in as a NEW profile (omni_ff) leaving the existing omni profile's behavior bit-identical; the checker (check_contract) was extended to force the feed-forward path to actually execute (cmd_vy=0.13) and promptly caught two genuine bugs: (1) re-clamping with soft_joint_pos_limits after the feed-forward (0.23 rad deviation) instead of reusing the parent's clip; (2) indexing processed actions by asset.joint_names instead of the action term's own contract-ordered _joint_names, which landed the feed-forward on the wrong joints (l_hip_yaw / r_ankle_pitch).
Change
Contract discipline as implemented: frozen dims + digest; manifest stamping and refusal; contract changes only via new versioned profiles; checker updated in the same commit as any pipeline change, with inputs chosen so new code paths are exercised.
Outcome
Both wiring bugs caught before any training or deployment ("两个都是 check_contract 当场抓出来的 —— 这次它值回票价"); old deployments provably unaffected by the new profile.
Mechanism
The contract is the only interface the policy and robot share; freezing plus fingerprinting makes divergence detectable, and an executable checker turns "the contract holds" from a belief into a test - but only if its inputs actually drive the new code path.
Applies when
- modifying the action or observation pipeline of a deployed policy
- exporting policies for hardware
- proposals that would change observation dims or history structure
“契约校验抓到的两个真错误（记账，别再犯）：1. 前馈后误用 soft_joint_pos_limits（URDF 限位 ×0.9）重钳 → 0.23 rad 偏差 … 2. 用 asset.joint_names 索引 _processed_actions → 前馈落到 l_hip_yaw/r_ankle_pitch 上 … 两个都是 check_contract 当场抓出来的 —— 这次它值回票价。”
train/C_LADDER_RUN.md § 3j. 契约级改动 / 契约校验抓到的两个真错误 A hardware run without its log is an anecdote - the first real get-up's policy, gain profile and log were never recorded, two CSVs stayed "to be reported", and runbook commands wrote different policies' logs under one copied filename
hardware-log-is-the-attribution-inputMake the log part of the run: name it from the policy and conditions automatically (never by hand-copied filenames), record the policy digest and gain profile inside it, include what the open questions need (torque, joint positions and targets), and treat a session without a collected log as incomplete.
Symptom
The recovery line's oldest open question - whether Isaac or MuJoCo reads torque demand correctly - was waiting on real-robot logs that never arrived, and the verdicts that did arrive could not be tied to files.
Context
deploy_policy writes a CSV per run (--log); the runbook's own analysis snippet reads its joint-position, target and action columns (q_, tgt_, act_), and the recovery hanging checklist asks for torque and joint logs for the whole run, to be compared with simulation. The first real get-up (08-09): policy, gain profile and log "to be recorded later". The first real A/B (08-11): v2_5b's result and both policies' CSVs "to be reported". In the runbook's walking commands, three runs of two different c4 policies log to real_s1e_pw08_teleop_0808.csv, and s1e and s2e_fric runs log to real_c2_700_pw08_teleop_0808.csv - filenames copied from other commands.
Change
None recorded; the spec kept listing the open-loop comparison as waiting for real logs.
Outcome
No real-robot log appears in the recovery spec through §50, so the simulator disagreement stayed unresolved and hardware verdicts stayed unattached to data.
Mechanism
Attribution needs the run's identity (policy digest, profile, conditions) and its signals in one artifact; a filename copied from another command mislabels the file, and a log not collected at the session is rarely collected later.
Applies when
- planning a hardware session whose result should settle a sim question
- log filenames are typed or pasted by hand
- hardware feedback arrives as prose without files
“python tools/deploy_policy.py --policy train/policies/omni_c4_ff800_pj.onnx … --log train/real_logs/real_s1e_pw08_teleop_0808.csv … q,t,a=d[:,c("q_")],d[:,c("tgt_")],d[:,c("act_")]”
RL系统/FOLLOW THIS copy 2.md § Walk 遥控 / S2 / csv 分析片段 (operator runbook, undated) Export every CAD part in the whole-machine frame so URDF rotations are zero and inertia is exact
urdf-shared-origin-exportGenerate the model so that correctness is structural: shared-origin STL export, zero rotations, subtraction-only origins, and an explicit 1e-9 g*mm^2 -> kg*m^2 conversion - never hand-rotate inertia tensors.
Symptom
Hand-assembled URDFs accumulate per-link rotation/origin errors and unit-conversion mistakes in inertia tensors - silent plant corruption that no later calibration can cleanly fix.
Context
Documented CAD -> URDF -> USD procedure from a successful Isaac Lab deployment, kept as the recipe if Lucen regenerates its model.
Change
(1) In CAD, align the whole robot to Z-up, X-forward (Isaac Lab convention) and ground the assembly; (2) export each STL with other parts hidden but the machine's shared origin kept, so all parts share one origin, every URDF rotation is 0, and inertia matrices equal CAD values directly; (3) units: Fusion 360 gives g*mm^2, URDF wants kg*m^2 - multiply by 1e-9; (4) link origin = negative of the joint position; link COM = CAD COM minus joint position; joint origin = difference of the two joint positions; (5) after URDF -> USD import, open the USD separately and set it instanceable before saving.
Outcome
A URDF whose rotations are all zero and whose inertia tensors are CAD-exact, eliminating an entire class of hand-transcription plant errors.
Mechanism
Keeping one shared origin turns every frame transform into a pure translation computable by subtraction, and leaves inertia tensors in the frame CAD already computed them in - no rotation of inertia tensors, the most error-prone manual step, is ever needed.
Applies when
- building or regenerating URDF/MJCF from CAD
- inertia or frame bugs suspected in the plant model
- importing URDF into Isaac Lab / USD
“导出 STL 时隐藏其他零件但导出整机——这样所有零件共享同一原点，URDF 里所有 rotation 全是 0，惯量矩阵直接等于 CAD 值 / 单位：Fusion 360 给 g·mm²，URDF 要 kg·m²，乘 1e-9 / link origin = 该关节坐标取负 … URDF → USD 导入后必须单独打开 USD 设成 instanceable 再存”
Experience.md § URDF 制作流程 (lines 87-92) Fine-tuning through a reward-table change was falsified (scatter, half-recover, collapse) - continuation training is legal only with the reward frozen
fine-tune-reward-change-falsifiedNever fine-tune through a reward-table change - retrain from zero; reserve checkpoint continuation for frozen-reward plant/DR widening, reset noise_std when branching, and watch for the scatter/half-recover/collapse signature as the abort trigger.
Symptom
The s1c A/B experiment: arm B fine-tuned from an existing checkpoint under the revised reward (same contract, same network, changed reward + small DR) and failed with a characteristic signature - scatter, half-recover, fall back ("打散→半恢复→摔回"); arm A trained from zero under the same config won decisively (full shaping lifted swing to 21.6 mm within 500 iters; shipped at 5500).
Context
Verdict recorded: "从零 + 强塑形是本机唯一验证过的发育路径" (from-zero plus strong shaping is this machine's only validated development path). The signature became a standing stop criterion in every later rung that touched a reward ("s1c B 臂签名,出现即停"). Crucially the boundary of the law was drawn explicitly when S2 continuation training was proposed: "当年证伪的是「奖励表中途改版的 fine-tune」… S2 续训奖励表全程冻结、只逐级加宽 plant/DR,属另一类" - continuing a checkpoint with the reward FROZEN while widening plant/DR one rung at a time is a different class and was allowed (and then worked, powering the whole S2/C lineage) - with the honest fallback that if frozen-reward continuation ever collapses, that rung retrains from zero and the doctrine gets re-examined with data. Fine-tune arms also need mechanical care: reset the checkpoint's collapsed noise_std (terminal 0.033 "会杀死探索") and account for iteration counters re-zeroing (curriculum gates fire immediately).
Change
Reward changes and lineage continuation permanently separated: reward revisions -> from-zero retrain; plant/DR widening -> frozen-reward continuation with per-rung gates; the B-arm signature promoted to a universal tripwire.
Outcome
No later reward revision was attempted by fine-tune; frozen-reward continuation carried S2 (PD/COM/friction rungs) and the C command ladder successfully from the s1e root.
Mechanism
A trained policy sits in an optimum of its reward's geometry; changing the reward moves the optimum but leaves the policy's exploration noise near-zero and its value function calibrated to the old returns - it disassembles the old solution faster than it can assemble the new one. Widening DR under a frozen reward instead keeps the optimum's identity and asks only for local robustification.
Applies when
- proposing to fine-tune an existing policy under a revised reward
- planning a robustification ladder from a validated checkpoint
- a continued run scatters then partially recovers then collapses
“B 臂 fine-tune 证伪(打散→半恢复→摔回——从零 + 强塑形是本机唯一验证过的发育路径)。… 当年证伪的是「奖励表中途改版的 fine-tune」(B 臂,塑形突变致终盘摔回);S2 续训奖励表全程冻结、只逐级加宽 plant/DR,属另一类;若 s2_lag1 续训本身塌方,回退方案 = 该级从零重训,续训教义再议(拿数据说话)。”
train/OMNI_V0_SPEC.md § 3. S1.3 / 4. 与 s1c fine-tune 证伪的关系 The hip_roll (l+r) asymmetry scalar predicted real-robot lateral drift - promote validated sim scalars into the gate
hip-roll-sum-predicts-lateral-driftHunt for cheap sim scalars that predict real-robot behaviors, validate them on direction AND ordering across multiple policies, then promote them into the acceptance battery; treat later violations as debt to justify in writing, not noise to ignore.
Symptom
A persistent hip_roll left/right asymmetry row in the sim2sim symmetry table had been dismissed as "calibration or mechanical asymmetry" noise; meanwhile real deployments drifted sideways by policy-dependent amounts.
Context
Forward-kinematics analysis reframed the scalar: both hip_rolls move the feet in +y for positive angle, so a same-signed (l+r) sum IS a lateral translation mode - the scalar is a direct lateral-drift bias estimate. Checked against real deployments: s1e (l+r = -0.0178, smallest magnitude) was the steadiest with least drift; 700 (+0.0253) drifted mildly left; A800 (+0.0267) drifted clearly left with the largest tilt 12.9 deg. Direction correct 3/3, ordering correct 3/3 (the log's heading calls it "四枚四中", four-for-four).
Change
The scalar was promoted into the acceptance battery as a posture-class criterion alongside tilt-max median: "hip_roll 左右不对称 |l+r| 不得比父代大" - doubling as a heat proxy (error ~ torque ~ heating).
Outcome
Used at every later gate; when the C4 product exceeded it by +0.005 rad (~+0.3 deg vs parent), the criterion was not silently waived - it was booked as explicit debt with a mechanism argument (the increment is task-required, far smaller than the sidewalk amplitude +/-2.2 deg) plus a related account (stand saturation 32.4% -> 37.2%).
Mechanism
A policy's static joint-angle bias in a translation-producing mode integrates into real-world drift; sim can measure that bias precisely and cheaply. A sim scalar earns gate status exactly when its predictions are validated against hardware in both direction and ordering - and a validated gate may only be exceeded with a written mechanism-level justification, never silently.
Conflicts
The log's heading says "四枚四中" (4/4) but the evidence table lists three policies and the text says "方向 3/3、排序 3/3"; the fourth instance is not shown in this file.
Applies when
- a real robot drifts or leans in a policy-dependent way
- deciding which sim measurements deserve gate status
- a validated gate criterion is marginally exceeded by a new product
“s1e | −0.0178（绝对值最小）| 微右、最不飘 | 三者中最稳、飘最小 ✓ … A800 | +0.0267 | 左、最飘 | 明显左飘、倾角最大 12.9° ✓ 方向 3/3、排序 3/3。 → 正式纳入验收表（与「倾角 max 中位」并列为姿态类判据）。”
train/C_LADDER_RUN.md § 3e. 顺带：hip_roll 左右不对称 (l+r) 就是横移偏置 —— 四枚四中 Raising a command bucket's share does not strengthen its per-state gradient - it only starves the other modes
bucket-share-is-not-a-gradient-leverWhen a skill is not learning, first prove its per-state signal is nonzero (ignore-floor and probe checks); only rebalance sampling shares to fix genuine sample starvation, and account the regression risk to the diluted modes before doing it.
Symptom
Sidewalk was not learning, and the reflex proposal was to give the side bucket a larger share of sampled commands.
Context
The C4-redo3 rung explicitly kept the 20/40/20/20 bucket (stand/forward/turn/side) with the reasoning written out: PPO computes advantages per state, so bucket proportion does not change the per-state gradient of side states; at 4096 envs x 20% x 24 steps the rollout already contained ~19.7k sidewalk states - sample count was not the bottleneck. And the cost side was already measured: cutting forward from 60% to 40% had made vx+0.30 die at +400 in an earlier run - more cuts would only collapse it sooner.
Change
Bucket proportions held constant across the entire C4 redo series; the actual bottlenecks (metric frame bug, reward variance penalty, exploration form) were pursued instead.
Outcome
Sidewalk was eventually fixed with zero bucket changes (feed-forward delivery, +100 iters); forward/turn skills never suffered starvation-induced regressions during the redo series.
Mechanism
Policy-gradient credit is assigned per visited state; oversampling a mode multiplies its states in the batch but not the informativeness of each, so if the per-state gradient is ~0 (behavior unreachable or reward indifferent), N times zero is still zero - while the displaced modes genuinely lose data and regress.
Applies when
- proposing to oversample a failing task/command mode
- a majority mode regresses after share rebalancing
- budgeting env count vs mode share for a multi-skill policy
“比例不动：PPO 逐状态算优势，桶占比不改变单状态梯度；4096 env × 20% × 24 = 每 rollout 已有 1.97 万个侧走状态，样本数不是瓶颈；而 forward 60%→40% 已实测让 f30 在 +400 处死掉，再加码只会更早塌。”
train/C_LADDER_RUN.md § 3i. 桶 20/40/20/20 不动（比例不动） A DR tail the robot never has is pure cost - stage deterministic plant levels instead of one wide uniform
dr-tail-plant-continuationSet every DR range from the measured deployment distribution and cut tails that hardware cannot produce; when an axis changes the controller's character (delay, major gain regimes), prefer staged deterministic levels with gates over one wide uniform.
Symptom
Two consecutive lineages (s1e, s1f) trained under uniform latency DR (0, 0.06 s = 0-3 frames) both converged to drag-glide gaits - buying survival under heavy delay by giving up swing (3.6 mm) - even though the real pipeline never exceeds ~2 frames.
Context
The account: roughly 1/3 of training quality was spent on the >2 frame tail that hardware never presents ("uniform 尾部 ~1/3 训练质量 花在真机不出现的 >2 帧上"). The deeper reading came from the user: uniform 0-3 frames is not merely tail-heavy - it "把性质不同的控制系统 混进同一 PPO batch" (mixes qualitatively different control systems into one PPO batch); a 0-frame and a 3-frame plant demand different controllers, and one policy trained on the mixture serves neither. The S2 v2 ladder therefore redefined latency "从「随机化参数」重新定 义为 actuator/control plant 的一部分": deterministic FIFO levels, staged 1 frame then 2 frames (lo=hi so fractional interpolation degenerates to exact N frames, synonymous with the harness --delay N), each level gated by the fixed acceptance battery - a plant continuation, not a randomization.
Change
Latency DR replaced by staged deterministic levels covering the measured 1-2 tick reality with no tail; each stage a separate continuation rung with the standard gate and rollback.
Outcome
The s2_lag1 rung showed the clean-signal benefit immediately (survival 20/20, heading 6x recovery) with the swing cost booked honestly (21 -> 12 mm, half-pass, ladder paused for adjudication); the drag-glide attractor from uniform tails did not recur.
Mechanism
DR asks one policy to cover a plant family; when part of the family is fictitious, the policy pays real capability for fictitious robustness, and when family members demand structurally different controllers, gradient averaging produces a compromise controller optimal for none. A measured, discrete plant set matches the actual deployment support and keeps each rung's training signal coherent.
Applies when
- policies converge to degenerate gaits that buy worst-case survival
- a DR range extends well past the measured hardware range
- choosing between wide randomization and a staged ladder on an axis
“两轮实证(s1e/s1f)宽尾延迟 DR 逼出拖地滑行 … uniform 0~3 帧不止尾重,而是把性质不同的 控制系统混进同一 PPO batch;1→2 帧确定性分级 = plant continuation,训练信号干净得多—— latency 从「随机化参数」重新定义为 actuator/control plant 的一部分。”
train/OMNI_V0_SPEC.md § 4. v2 阶梯 (2026-08-07 用户定) A constant-value plant rung passed every binary gate with record scores - and shipped 60% thinner posture margins that hardware exposed
constant-value-dr-overfits-marginRandomize deployment-critical axes over a narrow band spanning the measured real support - never a single value, never a fictitious tail - and report graded margin quantities (tilt margin) next to binary gates, because saturated gates rank thin-margin and thick-margin policies identically.
Symptom
s2_lag1 (trained at constant 1-frame latency) posted the strongest sim gate sheet in history (20/20 everywhere) yet was unstable on hardware, while s1e (trained across the full 0-3 frame band) was the every-run-stable SOTA at the same power.
Context
The sim autopsy (new --delay-jitter harness modeling the BusWorker's time-varying phase drift): 18 runs across constant and time-varying delays ALL survived - time variation alone does not kill - but the tilt-margin ordering reproduced hardware exactly: s1e 7.7-9.1 deg (thickest) < s2_lag1 10.7-15.0 < s1c 16.2-18.2. Attribution: constant-value training permits precise specialization to that one value; s1e's band diversity forced cross-value robustness - "恒定 1 帧训练 vs s1e 的 0~3 帧全带——分布多样性逼出跨值鲁棒,恒定值允许精确 特化" - so the constant-rung policy's margins were ~60% thinner, fine in sim's clean world, pushed over the line by real-world disturbances. Tool lesson booked: "存活门二值饱和后掩盖裕度差" - binary survival gates saturate and hide margin differences; graded margin columns (tilt-max) belong in the report. The synthesis with the opposite failure (wide tails cause drag-glide): the proposed resolution was a NARROW uniform band (0.02, 0.04) covering exactly the real 1-2 ticks - diversity inside the measured support, no tail, no single point. s1e's root selection later leaned on the same property: its full-band latency training "预装" the delay rungs and delivered "全工况稳定裕度" that survived power derating.
Change
DR-on-an-axis design refined to a three-way distinction: no wide fictitious tails (drag), no single constant values (thin margins), but a narrow band spanning the measured real support; acceptance reports gained graded margin columns alongside binary gates.
Outcome
The tilt-margin column entered the standard report; the s1e root (band-trained) carried the C ladder while the constant-value branch was archived with its three contributions credited.
Mechanism
Robustness margins are shaped by the diversity of the training distribution, not just its support: a point-mass distribution lets the optimizer trade margin for on-point performance, while a band forces solutions that keep margin across the band - and binary survival metrics cannot see the difference until the margin is spent on hardware.
Conflicts
The narrow-band (0.02,0.04) resolution was a pending recommendation ("裁决建议(待用户)") at the time of writing; the lineage instead moved root to s1e whose full-band training predated the staged ladder - the deterministic-staging card and this card record the two failure modes the final design must avoid simultaneously.
Applies when
- a rung trained at a fixed plant value aces sim but wobbles on hardware
- binary acceptance gates are all saturated across candidates
- choosing between constant, banded, and wide DR on one axis
“18 跑全活,时变性单独不足以击杀;但 tilt_max 裕度排序完整复现真机:s1e 7.7~9.1°(最厚)< s2_lag1 10.7~15.0 … 恒定 1 帧训练 vs s1e 的 0~3 帧全带——分布多样性逼出跨值鲁棒,恒定值允许精确特化 … 存活门二值饱和后掩盖裕度差(s2_lag1 sim 门 20/20 史上最强却真机不稳)”
train/README.md § s2_lag1 真机不稳 × s1e 稳的 sim 对拍(2026-08-07,时变延迟实验) action_rate weight is the sim2real bandwidth knob - re-tune it whenever a rate limiter is removed
action-rate-weight-vs-bandwidthSet action_rate weight relative to real actuator bandwidth, and re-tune it any time another smoothing/limiting element (filter, slew limiter, gain) changes - reward weights are load-bearing parts of the actuator model.
Symptom
With a low action_rate_l2 weight the policy learns fast actions; the unmodeled part of the actuator response is then excited hardest, and sim2real "直接崩" (collapses outright). With too high a weight, actions become so slow the robot cannot maintain balance.
Context
The reference developer called action_rate_l2 the single most important reward for transfer, with side-by-side video evidence that the high-penalty, slower policy is clearly better on hardware. Lucen context: the team had just removed the SOFT_SPD=1.0 velocity limiter, which had been an implicit actuator-bandwidth constraint - leaving action_rate as the only remaining constraint on action speed.
Change
Decision recorded: after removing SOFT_SPD, re-evaluate the action_rate weight rather than keep the old value, since its effective role changed from "additional smoother" to "sole bandwidth constraint".
Outcome
Logged as a priority follow-up ("重新评估 action_rate 权重 - 拆掉 SOFT_SPD 之后这一项的作用变了"); the failure mode it guards against is training high-frequency actions the real actuators cannot track.
Mechanism
Slower actions stay inside the frequency band where the ideal-PD sim actuator and the real actuator agree; fast actions probe the band where unmodeled delay, inductance, and bandwidth limits dominate, so model error is amplified in exact proportion to action speed. Any removed external rate limit transfers that constraint's entire job onto the action_rate penalty.
Conflicts
The low/high tradeoff evidence is the external developer's report (with video); the Lucen-side entry is a pre-registered risk and decision, not yet an on-robot A/B at the time of writing.
Applies when
- removing or adding an action filter, slew limiter, or low-level speed cap
- real robot shows high-frequency chatter or overheating absent in sim
- tuning smoothness rewards before a hardware deployment
“权重低 → 动作快 → 执行器模型不准的部分被放大，sim2real 直接崩 / 权重高 → 动作慢 → 好迁移，但可能慢到无法维持平衡 … 我们刚拆掉 SOFT_SPD=1.0 的限速器，等于把执行器带宽约束整个移除了。action_rate 惩罚现在是唯一还在约束动作速率的东西，需要重新评估权重”
Experience.md § action_rate_l2 是他认为最关键的 reward (lines 61-70) Three same-shaped judging errors - task metrics (survival, tracking, displacement) cannot stand in for posture metrics
task-metrics-vs-posture-metricsKeep validated posture-class rows (tilt, per-joint L/R asymmetry, temperature) in every acceptance battery alongside task rows; when operator feel contradicts the gates, suspect the metric class before the operator - and never build a new skill on what is actually an asymmetry defect.
Symptom
The C2 product judged "full pass" on task metrics (A800: turn-gap 7 pp, vx+0.30 19/20) felt WORSE in the operator's hands than the half-pass 700: A800 tilted up to 12.90 deg (700: 6.64), drifted left while standing, showed larger per-joint asymmetries, and ran its hip_rolls 5 degC hotter.
Context
The re-judgment catalogued three same-type metric errors in one campaign: (1) stand judged by SURVIVAL - missed 0.5-1.4 m wandering; (2) stand ranked by DISPLACEMENT - ordering was opposite to real feel (tilt ordering matched); (3) chirality judged by wz-tracking GAP - measured turning symmetry while the robot's actual disease was postural left/right asymmetry, "两个不同的东西,且结论相反". Common pattern named: "我一直用「任务指标」当判据,而真机手感对应的是「姿态 指标」… 任务类指标不能替代它". The fix was already in the data: the per-joint left/right asymmetry table (printed identically by sim2sim and deploy) agreed with hardware in direction on every row - "判据可用、有预测力,我只是没把它写进 PASS 条件". Shipping decision followed the posture read: product reverted to 700 ("又一次「买到 精度、卖掉别的」"), and the C4 root moved to 700 as well, with the sharpest line of the episode: A800's left-drift "like sidewalking" is probably its frontal-plane asymmetry defect, not a capability - "在缺陷上建能力是危险的".
Change
Two posture quantities with demonstrated real-robot predictive power promoted into every PASS battery: tilt-max median and per-joint left/right asymmetry (both sim-computable, deploy-homologous); motor-temperature readout added to session close-out.
Outcome
Deployment flipped to the posture-better checkpoint; the hip_roll temperature table (43-48 degC vs 25-28) confirmed the earlier 90%-of-heat account; the run-line acceptance battery inherited the posture rows from birth ("任务类替代不了姿态类").
Mechanism
Task metrics measure goal attainment under the evaluator's episode definition; posture metrics measure the body state trajectory that operators, motors, and downstream skills actually experience. The two can rank candidates oppositely because task success tolerates postural pathology - so a battery without posture rows is blind to exactly what hardware feel reports first.
Applies when
- hardware feel disagrees with a green acceptance table
- choosing between checkpoints that split task vs posture metrics
- selecting the root for a skill that resembles an existing defect
“共同模式:我一直用「任务指标」(存活 / 跟踪率 / 位移)当判据,而真机手感对应的是「姿态指标」(倾角、逐关节左右不对称)。→ 验收判据里必须有姿态类指标,任务类指标不能替代它。… A800 的「左飘像 side walk」很可能 … 是它更大的额平面不对称的表现 —— 在缺陷上建能力是危险的。”
train/README.md § C2 选点改判 (2026-08-09): 手性判据第三次选错指标 Four in-lineage attempts to widen the standing stance failed - remove a tax, add a joint-space knife, change the target, add a task-space metric penalty - because the stance was the end state of the get-up path; trained from scratch with the right terms it grew right from day one
stance-decided-by-get-up-pathA posture a skill ends in is shaped by the path the policy takes to reach it; if several single-variable edits to the terminal-phase reward cannot move it, stop editing that phase and retrain with the terminal constraint present from the start.
Symptom
v2_6c stood with its feet 0.159 m apart (task-space) and its hips yawed 45-47 deg the same way, which split on the real robot. Standing-phase reward edits did not move it.
Context
V2.7-A removed the flat-feet tax on compensated stances (stance unchanged); V2.7b added a hip-roll lower-bound hinge (+5 deg in 3,000 iterations, yaw ratchet); V2.8 changed the stand_pose target to a wide flat stance (stance unchanged, yaw not unwound, feet nearly overlapping, mu 0.4 transfer 2%); V2.9 penalized lateral spacing in metres (the policy parked just outside the penalty's gate in a lunge, 0% success). The v2_6c get-up goes through a split and closes the feet together as it rises.
Change
In-lineage stance surgery was formally closed. V3.1 trained from scratch with task-space stance terms present from the first iteration (and, after P1, a positive width band instead of a penalty).
Outcome
V3.1 P1b: lateral stance 0.364 m, foot tilt 0.0 deg, all four categories 100%, MuJoCo mu 1.0 and 0.4 both 100% - with a symmetric toe-out the kinematic audit had not enumerated. P1c (with a yaw guard): 0.355 m, all six acceptance criteria passing, mu 1.0-0.4 all 100%; it became the product.
Mechanism
A converged policy does not rebuild the path that produced its terminal posture; a standing-phase gradient only finds the nearest hack around the posture the get-up delivers.
Applies when
- the final posture of a transition skill is wrong and resists terminal-phase shaping
- repeated continuation rungs produce hacks instead of the intended posture
- deciding between another in-lineage fix and a from-scratch retrain
“窄站距 + yaw 扭是 v2_6c 起身策略(劈叉起身 → 双脚并拢收势)的**结构性 终态**,不是站立段的孤立参数 —— 站立形态由起身路径决定,在血统内只动 站立段奖励改不动它。”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §44 结果:V2.8 判 FAIL —— 血统内站姿手术第三次证伪 Record-high training reward hid a fully-failing DR subgroup - aggregate metrics average over draws, gates must test per condition
aggregate-metrics-mask-subgroup-failureNever gate on metrics aggregated across DR draws: evaluate at fixed representative conditions (especially the deployment-critical stratum), and if a difficulty axis matters, ramp it on measured per-stratum success rather than sampling the full range from iteration zero.
Symptom
omni_s1e trained under constant-wide latency DR (0, 0.06 s) posted the lineage's highest-ever Isaac reward (129) - while the --delay 2 smoke evaluation showed 3/3 falls from iter 1500 onward, persisting to early stop; the usable checkpoint window shrank to iters 500-1000.
Context
Diagnosis written plainly: "聚合奖励掩盖重延迟尾部子群体失败" - the aggregated reward averages over latency draws, so the majority of light-delay environments can mask the total failure of the heavy-delay tail. The remedy for the training side was a survival-gated ratchet curriculum (survival_gated_latency): the sampling cap starts at 0.02 s and rises +0.01 only when a 4096-reset window's survival (time_out share) reaches >=90%, capped at 0.06, ratchet up-only - "增益与延迟耐受一起长,不升到策略撑不住的 地方" (gain and delay tolerance grow together; never raise past what the policy can hold). The detection side was already in place from the noise-crutch episode: the per-condition smoke curve, not the training reward, is the health readout.
Change
Latency exposure made curriculum-gated on measured subgroup survival instead of uniform-from-zero; per-condition (--delay 2) smoke evaluation kept as the authoritative curve; watcher scoring adjusted (survival weighted 3x) so recovery during hard phases is not early-stopped away.
Outcome
The failure mode was caught by the smoke curve within one generation; the follow-up redesign (deterministic staged latency) superseded the ratchet, but the aggregate-masking lesson held through both.
Mechanism
Expected-return training weights each DR draw by probability, so a subgroup can contribute bounded loss while being catastrophically failed; any scalar averaged over the randomization cannot distinguish "uniformly decent" from "great on easy draws, dead on hard ones". Only conditioning the evaluation on the stratum reveals the split, and curricula should raise difficulty on measured stratum success, not on schedule.
Applies when
- training reward hits records while a fixed-condition eval degrades
- wide DR on an axis where deployment sits at one known value
- designing curricula for difficulty axes (delay, push, terrain)
“常量 latency DR (0,0.06) 从零训被证伪——Isaac reward 129 历代最高,但 --delay 2 冒烟 iter1500 起 3/3 全摔持续到早停(聚合奖励掩盖重延迟尾部子群体失败,可用窗口只剩 500/1000)。… 采样上限 0.02 起步 … ≥90% 才 +0.01s,0.06 封顶,棘轮只升不降。”
train/OMNI_V0_SPEC.md § 3. S1.5(s1e 训练塌方复盘) A time gate that had cured one lineage's rushing made the from-scratch lineage trade away its stance width twice (0.364 -> 0.235 m, 0.355 -> 0.251 m) - its disease was absent there, so the fix was retired and the pre-gate checkpoint shipped
time-gate-vs-wide-stance-retire-the-fixCarry a fix into a new lineage only if its disease is present there; a mechanism that cured one lineage can be net negative in another, and when doubling a term's weight recovers almost nothing, treat the two objectives as structurally in conflict and remove the one whose purpose is gone.
Symptom
V3.1's phase 2 (the 3 s zero gate on standing income, continued from P1b) kept 100% success on every friction level and slowed the get-up, but the lateral stance drifted 0.364 -> 0.235 m and hip yaw crept to 57 deg against its 60 deg limit. With the width band's weight doubled (P2c, after P1c) it drifted again, 0.355 -> 0.251 m, below the pre-registered 0.30 m failure line.
Context
The zero gate had been introduced in V2.5/V2.5b to slow the old lineage's get-up. In V3.1 the rushing was already absent: P1c got up in 0.90-1.06 s with a worst torque ratio of 73.3%, better than the stamped v2_6c, because the full beta curriculum, second-difference smoothing and pull curriculum had cured the violence inside training.
Change
Recorded as a candidate law with two data points - the zero gate and a wide stance are mutually exclusive here - and the zero gate was removed from the V3.1 recipe. P1c (the pre-gate checkpoint) went through the full stamp-level acceptance instead.
Outcome
P1c passed everything: all six criteria, lateral stance 0.355 m, foot tilt P75 2.0 deg, mu {1.0, 0.8, 0.6, 0.4} x 10 seeds all 100%. recovery_v3_1p1c.onnx was stamped and pushed to the robot channel.
Mechanism
The zero gate moves the income toward "stay stable until the end", and under low-friction DR a wide stance has a slip tail, so survival outbids the width band; doubling the band's price bought back only 0.016 m - an auction that does not converge signals structural conflict, not an under-priced term.
Applies when
- porting reward mechanisms from an old lineage into a fresh recipe
- a width, margin or posture metric erodes during a late training phase
- a weight increase produces a negligible change in its target
“**定律候选(二实证):归零门 × 宽站互斥**。 … 加价翻倍只挽回 0.016,竞拍不收敛)。 … **归零门是 v2_5 血统的历史包袱,对 V3.1 配方是净负资产,P2 阶段除名 —— P1c 即终点形态**。”
git:Lucen-recovery@origin/recovery:train/RECOVERY_V0_SPEC.md § §49 终验(2026-08-14) When the eval proxy has a known systematic bias, gate on within-proxy differences, not absolutes
relative-metrics-survive-proxy-biasWhere the evaluator is known-biased, design gates as within-evaluator contrasts (left vs right, A vs B, pre vs post) that cancel common-mode error; reserve absolute thresholds for quantities whose proxy calibration has been checked.
Symptom
MuJoCo systematically overestimated yaw turn gain (1.58/2.45 where Isaac measured 1.04/1.02), and friction alignment recovered only part of the gap (mu 1.0 -> 0.6 pulled it to 1.43/2.27) - so any absolute turn-gain acceptance threshold in MuJoCo would be judging the proxy's bias, not the policy.
Context
The acceptance criterion was rewritten to use only the left/right difference of the turn gain (standard: <20%): both directions pass through the same biased proxy, so the bias largely cancels in the difference while the policy's chirality - the thing being gated - survives. The absolute-gain row was dropped: "转向只判左右差,不判绝对值". A parallel task was still opened to align MuJoCo contact parameters to Isaac's (mu, restitution), since drift proved highly friction-sensitive (straight-line drift -65 deg at mu 1.0 vs +1 deg at mu 0.4) - bias reduction and bias-robust metrics proceeded together.
Change
Gate metric changed from absolute turn gain to left-right gain difference; proxy-alignment work scheduled separately rather than blocking acceptance.
Outcome
Turn acceptance became meaningful across proxy versions (v5 70% -> v6 19% difference measured the real improvement) while the absolute bias question was pursued without holding the ladder hostage.
Mechanism
A systematic multiplicative or additive proxy bias applies to both arms of a mirrored measurement; differencing (or ratioing) mirrored conditions cancels the common-mode bias to first order, leaving the asymmetry signal. Metrics built this way remain valid while the proxy is imperfect - which it always is somewhere.
Applies when
- a sim proxy disagrees with the trainer or hardware in absolute terms
- writing acceptance thresholds for direction-paired skills
- proxy calibration work would otherwise block a ladder
“转向只判左右差,不判绝对值 —— MuJoCo 的偏航增益系统性高估(实测 1.58/2.45 vs Isaac 1.04/1.02),摩擦只能解释一部分(μ 1.0→0.6 仅拉到 1.43/2.27)。差值是相对量。”
train/WALK_V5_SPEC.md § 6. 验收 Fix the task first, harden the plant second - DR budget spent on a dying task is wasted
task-shaping-before-plant-hardeningFreeze the task/command distribution before spending DR budget on plant robustness; if the task will still change, schedule plant hardening as a final pass and book the interim robustness gap explicitly.
Symptom
Tempting default ordering was to keep the plant-hardened (S2) lineage and teach it new commands; but the S2 plant adaptation had been earned on the straight-walk task, and the new omni tasks (sidewalk, in-place turn) use completely different contact patterns.
Context
The team had direct evidence that DR robustness is a budget that gets reallocated when the data distribution changes ("push/μ 两轮已实证 DR 预算有限且会被重分配") - robustness trained under one task/command distribution does not persist when training continues under another.
Change
Ladder order set to: first C (task shaping - add command modes until the task family is final), then a second S2 pass (plant hardening) on the C product. The plant-robustness gap this creates mid-ladder is accepted and booked explicitly ("此处不欠账" - the debt is assigned to the second S2 pass, not denied).
Outcome
The first S2 pass was not wasted: its laws (kd bandwidth <-> low mu, push need not be trained, ground mu need not be trained, bistability) let the second pass drop from five rungs to three. The C ladder itself ran on the softer plant band without incident.
Mechanism
DR robustness is carried by the policy's visited-state distribution; changing the task changes that distribution, so robustness bought under the old task partially dissolves. Hardening before the task is final means paying for robustness on states that will no longer be visited - "给一个即将不存在的任务花预算" (spending budget on a soon-to-not-exist task).
Applies when
- deciding ordering between skill/command expansion and DR hardening
- a hardened lineage is proposed as the root for a task change
- robustness regressions appear after adding new command modes
“S2 的 plant 适应是为直行步态调的，C4 侧走/C3 原地转是完全不同的接触模式，先硬化再改任务 = 给一个即将不存在的任务花预算（push/μ 两轮已实证 DR 预算有限且会被重分配）。故顺序改为 先 C（任务定型）→ 再 S2（plant 硬化）。”
train/C_LADDER_RUN.md § 0. 决策逻辑 = 短板可不可恢复 (末段) Compare the achieved reward to the computed ignore-floor to tell "never learned" from "learned but unprofitable"
ignore-floor-diagnosisFor any skill that trains flat, compute the reward the null policy would earn on that term; achieved==floor means the behavior never paid out (find why: exploration, reward observability, or feasibility) - do not tune weights first.
Symptom
C4 sidewalk failed on both arms; the question was whether the policy had found sidewalk and rejected it as unprofitable, or never found it at all - two diagnoses with opposite fixes.
Context
The theoretical value of the sidewalk tracking term for a policy that completely ignores the command was computable from the command distribution: 0.189. Trained final values landed at 0.187 (arm A) and 0.204 (arm B) - sitting exactly on the ignore-floor - while the honest balance at the optimum actually favored sidewalking (net +0.88/step inside the side bucket). Later the same arithmetic closed the whole saga: for the true reward landscape, doing real sidewalk scored 0.153 vs 0.944 for ignoring - the policy's refusal "是理性最优，不是探索失败" (rational optimum, not exploration failure) under one hypothesis, and under the final measurement-corrected account the policy had "每一次都在 做理性选择" (made the rational choice every time).
Change
Diagnostic rule adopted: compute the ignore-floor for the new term; if the achieved value sits on it, the behavior was never expressed in useful volume (or the reward cannot distinguish it - check both); if the achieved value is above floor but the behavior is absent at deployment, the policy sampled it and priced it out - then the reward balance, not exploration, is the lever.
Outcome
Correctly identified that PPO had not merely under-valued sidewalk; each subsequent hypothesis (waveform sign, regularization cage, exploration form, reward kernel) was tested against this floor arithmetic, which kept the search honest through three reversals.
Mechanism
Every reward term has a computable value under the null behavior; the achieved-vs-floor gap is a one-number audit of whether the optimizer ever monetized the target behavior. It converts "training failed" into one of two mechanistically distinct states with different fixes.
Applies when
- a new skill's tracking reward plateaus early
- deciding between exploration fixes and reward-weight fixes
- post-mortem of a failed curriculum rung
“track_lin_vel_y_exp 训练终值恰好坐在「完全无视指令」的底分上（A 0.187 / B 0.204，理论值 0.189），而终点 balance 明明有利（side 桶内净 +0.88/步）。不是学会了不划算，是根本没学到。”
train/C_LADDER_RUN.md § 3e. 为什么首战 FAIL Adapting a lineage to one plant increment needs hundreds of iterations, not thousands - long runs only buy specialization
continuation-budget-not-from-zeroBudget continuation rungs by increment class (hundreds of iterations for plant pins and smooth shifts, ~1000-1500 only for behavior-demanding changes like push), enforce a hard cap with frequent evaluation, and treat remaining budget as a reason to stop, not to continue.
Symptom
The default "6000 iterations per rung" (a from-zero-scale budget) was about to be applied to continuation rungs whose only change is one plant/DR increment - overspending compute and, worse, giving each rung thousands of iterations to specialize away retained skills.
Context
The 2026-08-07 budget table replaced the default with "最低适应窗口 + 每 100 iter 验收 + hard cap" scaled to the increment's difficulty: fixed-latency levels 300-500 (cap 500-800; the base has already seen in-band values, this only pins the plant); PD full-band 700 (cap 1000; kp+/-20%/kd+/-30% clearly widens the actuator family); COM +/-20 mm 500 (cap 800; a smooth dynamics shift); friction DR 700 (cap 1000; contact and actuator friction change the gait/contact solution together); push 1000 (cap 1500; a non-static plant change requiring recovery behavior - hardest). Rationale: "续训适应一个 plant 增量不需要从零量级的预算,跑长了只是给特化时间". The C ladder reused the scheme (per-rung caps 500-2000 by increment type), and the deep-training hazard got its own name when long runs sold quality ("深适应卖质量" - deep adaptation sells quality).
Change
Per-rung iteration budgets set by increment class with hard caps and 100-iter watch loops; checkpoint selection inside the window by the smoke curve, never "run to cap because budget remains".
Outcome
S2/C rungs completed in 300-1500 iterations each; the recurring late-run degradations (collapse valleys at 1500+, vx+0.30 decay) fell outside most rungs' caps instead of inside their runs.
Mechanism
A continuation rung's learning problem is local robustification around an existing optimum - low sample complexity; iterations past adaptation are spent sharpening onto the current distribution, which is exactly how retained skills and margins erode. Budgets sized to the increment bound both compute and the specialization damage window.
Applies when
- planning iteration budgets for a robustification or command ladder
- a continuation run keeps improving its training metric late
- retained skills decay in the back half of long continuation runs
“「最低适应窗口 + 每 100 iter 验收(watch_ckpt --every 100)+ hard cap」—— 续训适应一个 plant 增量不需要从零量级的预算,跑长了只是给特化时间 … ⑥ push | 1000 | 1500 | 非静态 plant 变化,要学 recovery 行为,最难”
train/OMNI_V0_SPEC.md § 4. 每级 iter 预算(2026-08-07 用户定) Every rung gets written stop criteria - hit any one, stop; tightened when priors say results should come fast
preregistered-stop-criteria-per-rungFreeze per-rung stop criteria (old-skill floors, new-skill deadline, oscillation signature, known pathology signatures) before training, stop on first hit - and shorten the deadline in proportion to how fast your mechanism says results should appear.
Symptom
Continued training past the point of degradation had previously destroyed capabilities (C1 trained away the root's backward skill); without hard stop rules, sunk-cost reasoning keeps runs alive too long.
Context
Standard rung protocol - eval_c_matrix every 100 iters at 5 seeds with a fixed criteria list, frozen before training ("开训前写死，事后不许改"): (1) stand or vx+0.15 at <=3/5 for two consecutive points; (2) vx+0.15 tracking <70% for two consecutive points (vs recorded root baseline); (3) oscillation signature - survival repeatedly crossing zero between adjacent checkpoints, stop on first occurrence, do not wait for confirmation; (4) new-skill-no-progress deadline (e.g. wz tracking <30% or still same-signed after iter 400); (5) a known lineage pathology signature (s1c arm-B: scatter -> half-recover -> collapse).
Change
Stop budgets scale with prior knowledge: when C4-redo4's feed-forward was already proven open-loop, the no-progress deadline was tightened from +500 to +100 iters ("前馈是开环就给 71~96% 的 … 若 +100 还没有 … 不值得再烧").
Outcome
Multiple rungs were stopped exactly on criterion (C4-redo criterion 4 at iter 1200; redo2/redo3 at absolute 900), converting each into a clean hypothesis test instead of a drifting run; no rung burned its full cap on a dead hypothesis.
Mechanism
Degradation of retained skills is a lineage-level injury that later training often cannot undo, so detection latency is capital loss; and a stop rule written before the run cannot be bent by hope. Tightening deadlines when priors predict immediate results converts "no result yet" into evidence against the mechanism.
Applies when
- starting any resumed/curriculum training rung
- deciding whether to keep training a run that shows early regression
- a mechanism-backed change should produce results immediately
“每 100 iter eval_c_matrix.py --seeds 5，命中任一即停 … 存活在相邻 checkpoint 间反复跨越 0（出现即停）… 收紧版：绝对 iter 900（= +200）时 vy±0.10 跟踪仍 <30% 即停”
train/C_LADDER_RUN.md § 3h. 预注册停梯判据（开训前写死，事后不许改） One fixed acceptance matrix for every rung - new skill must PASS while every old skill stays within a regression budget
fixed-acceptance-matrix-per-rungFreeze one acceptance matrix for the whole ladder; every promotion requires the new skill's PASS plus bounded regression on every prior skill, measured against the parent's baseline under the current (bug- fixed) metric code - and include command transitions, not just steady states.
Symptom
Sequential skill training silently trades old skills for new ones (C1 trained away the root's backward ability); without a constant measurement frame, each rung's numbers are incomparable and regressions hide.
Context
The C ladder ran the same 13-cell command matrix at 20 seeds per cell at every rung (stand; vx +0.15/+0.30; vx -0.10/-0.20; vy +/-0.10; wz +/-0.20; two vx&wz combos; two vx&vy combos), with promotion requiring "新技能 PASS 且旧技能不明显退化" - old-skill regression budget <=2/20 against the parent's recorded 20-seed baseline. For the transition-rich final rung, a command-switch block was added (forward->stop, stop->backward, forward->turn, left->right, turn->forward; survival + re-track within 2 s) because "每个 steady command 都会做 ≠ 命令切换不会摔" - steady-state success does not imply switch safety, and the joystick does switches. The C4 product's gate ran 260 cells (13 x 20) all 20/20.
Change
Battery frozen once, reused verbatim per rung; baselines re-measured per parent (and re-measured again after the metric-frame fix, since old baselines were taken with the buggy coordinate reading - "旧基线是坏坐标系的， 不可引用").
Outcome
Regressions were caught at the rung that caused them (C1's backward loss, C2's vx+0.30 decay), and cross-rung comparisons stayed valid for the ladder's whole life.
Mechanism
A constant matrix makes every rung's output a point in the same metric space, so "did we lose anything" is a table diff, not a judgment call; the per-skill regression budget converts previously earned PASSes into standing constraints on all future training.
Applies when
- designing gates for sequential skill addition
- promoting a checkpoint to be the next rung's root
- after any evaluation-code fix (old baselines must be re-measured)
“新技能 PASS 且旧技能不明显退化才晋级。… C5 追加：命令切换验收（steady ≠ transition） forward→stop、stop→backward、forward→turn、left→right、turn→forward，各 20 seed，判存活 + 切换后 2 s 内是否重新跟上。”
train/C_LADDER_RUN.md § 5. 固定验收矩阵（每级跑同一张，每项 20 seed） A sim veto needs real confirmation too - the worst sim cell was scheduled as the most informative hardware run
sim-veto-needs-real-confirmationNever let sim alone both condemn a purpose-built configuration and escape audit: spend one cheap, safeguarded hardware run on the condemned cell, pre-registering what agreement and disagreement would each imply about the proxy.
Symptom
The fric-2400@kd1.0 combination was sim's worst cell across the board (survival 17/20 - the only miss, mu0.4 1/20, push 103/160, zero-cmd 2/20), yet it was the only product specifically trained for the kd1.0 deployment gain - discarding it on sim evidence alone would leave the sim's own validity untested exactly where it mattered.
Context
The team had been burned in the other direction before ("Isaac 指标三次 零预警" - training-side metrics gave zero warning three times), so the symmetric rule was written: sim's rejection also needs hardware confirmation ("sim 判被支配 ≠ 真机被支配 … sim 的否决也要真机确认"). The run was pre-registered with a dual reading: real matches sim -> the S2f ladder closes and the fork root is settled; real clearly better than sim -> the MuJoCo proxy has a systematic bias in the kd1.0/low-margin region, "那比选型本身重要得多" - and every S2f sim acceptance would need re-scoring.
Change
The condemned configuration was kept on the hardware roster (last, spotted, minimal exposure) explicitly as a proxy-validation probe, not as a deployment candidate.
Outcome
Session design captured either result as progress: selection confirmed, or a proxy bias discovered that would re-price the whole ladder's verdicts.
Mechanism
Every sim verdict is a joint statement about the policy AND the proxy; cells where a policy was purpose-trained for the exact condition sim condemns are where proxy error is most likely and most costly. Testing the veto converts a selection decision into a calibration measurement of the evaluator itself.
Applies when
- sim rejects the configuration that targets the actual deployment condition
- the eval proxy's calibration has never been checked in that regime
- deciding which hardware runs are worth their risk
“但它也是唯一为 kd1.0 部署档专门训的产物 —— sim 判被支配 ≠ 真机被支配, 「Isaac 指标三次零预警」的教训反过来同样成立: sim 的否决也要真机确认。… 若真机明显好于 sim → MuJoCo 代理在 kd1.0/低裕度区有系统性偏差, 那比选型本身重要得多。”
train/REAL_RUN_S2.md § 上机名单 note / 2. sim 侧预注册预期 ⑤ mj_objectVelocity returns inertial-principal-axis frame - one API assumption poisoned eval and observations for a whole line
body-frame-velocity-api-auditVerify every frame-sensitive API against a hand-computed truth (rotate raw qvel yourself, or command a known world velocity and check where it lands) before trusting any evaluation or observation built on it - especially when a model's inertial frame is rotated from its body frame.
Symptom
Sidewalk vy read ~0 under every condition; separately, whole-policy performance was mysteriously mediocre in sim2sim while training-side numbers looked fine. Four training rungs were declared FAIL partly on these readings.
Context
base_link's URDF inertial frame is rotated 90 deg about x relative to the body frame (iquat = [0.7071, 0.7071, 0, 0]). mj_objectVelocity(flg_local=1) rotates into ximat - the inertial principal-axis frame - not the body frame, and returns center-of-mass point velocity, not body-origin velocity. Consequences measured: the "vy" column was actually vertical velocity vz (walking at cmd 0.25: old reading +0.0093 vs true -0.0424); the angular velocity fed to the policy in sim2sim was [wx, wz, -wy] - a different quantity than Isaac and the real IMU provide. RMS check over 8 s of walking: y/z axes swapped between v6[:3] and the qvel truth.
Change
Fixed sim2sim and both probes to compute ang_b = qvel[3:6] and lin_b = xmat.T @ qvel[0:3] (identical quantity to Isaac's root_ang_vel_b / root_lin_vel_b), with a standalone reproduction script (frame_bug_repro_0809.py).
Outcome
Re-scoring the "failed" C4 lineage under correct coordinates reversed the verdicts: c4r4 checkpoints showed vy 80-126% tracking (old reading: +/-2%) and vx+0.30 at 91-95% where the old metric said 0/5 - the bad frame both mis-measured vy and, via corrupted policy observations, systematically depressed all measured performance. Final product passed 260/260 cells.
Mechanism
A simulator API's frame convention is part of the observation contract; when the model's inertial frame is rotated relative to the body frame, frame-agnostic use of a "local" velocity silently permutes axes. Feeding a policy an axis-permuted angular velocity is an observation corruption that degrades behavior everywhere, not just on the axis being studied.
Applies when
- building or auditing a cross-simulator evaluation harness
- one measured axis reads near-zero under all conditions
- sim2sim scores are inexplicably worse than training-side metrics
- URDF/MJCF inertial frames are rotated relative to body frames
“base_link 的 iquat = [0.7071, 0.7071, 0, 0] … mj_objectVelocity 用的是这个 … 喂给策略的 base_ang_vel 是 [wx, wz, −wy] —— MuJoCo 侧观测与 Isaac / 真机 IMU 不是同一个量；vy_mean 报的是竖直速度 vz —— 前进 cmd 0.25 时旧读数 +0.0093，真值 −0.0424。”
train/C_LADDER_RUN.md § 3l. ⚠️ mj_objectVelocity 读的是惯性主轴系 / 3m. 一 bug 坐实 Training at scale 0.4 is NOT the twin of deploying 0.5 at power 0.8 - the algebra matches, the learned policy does not
deploy-scaling-not-training-equivalentNever assume deploy-side scalings can be folded into training-time constants ("burning the crutch into training"): the learned optimum depends on the training-time authority, so treat such conversions as full experiments with pre-registered expectations and a sim2sim gate before any hardware.
Symptom
s1g (S1.6) trained from zero at action_scale 0.4 - meant as the "training twin" of the hardware-proven s1c-at-power-0.8 (0.8 x 0.5 = 0.4) - was all green in Isaac (zero falls, reward 117) yet scored 0/3 across all eight checkpoints and 0/20 at 20 seeds in the MuJoCo gate, falling forward at median 1.57 s with a 2.9x speed overshoot.
Context
The pre-registered expectation (survival gate should pass, since the conviction matrix showed s1c@0.8+delay2 all-survive) was cleanly falsified, and the harness was acquitted by controls: --delay 0 fell identically (not a delay fragility), check_contract all green, and s1c through the same harness survived 2/3. The verdict: "「s1c@0.8 = 0.4 训练孪生」的代数等价不成立" - a policy deployed with a derated output still LIVES in the 0.5 internal model it trained under (its value function, its expectations of its own authority), while a policy that starts training with reduced authority learns a different, clip-hugging gait with zero margin for plant differences ("部署端打折的策略活在 0.5 的内模里,训练起点收权限学出的是贴 clip 的 另一套步态,对 plant 差异零余量"). Result: the policy was withdrawn before hardware ("撤回——不上真机"), the lineage root moved back to the 0.5-contract s1c-5500, and this became the C ladder's cited fact-check ("s1g 是 0/20 证伪出局的那一代").
Change
The amplitude-surgery route abandoned; contract kept at scale 0.5; the deploy-side 0.8 crutch later retired on its own merits when the delay-complete s1e generation ran at full power.
Outcome
One training run bought a clean falsification of a plausible algebraic identity; no hardware time was spent on it because the sim2sim gate caught it.
Mechanism
Output scaling commutes with the network arithmetic but not with learning: the training-time scale shapes which gait solutions are reachable and how much clip headroom the optimum keeps. A derated mature policy retains the wide-authority solution executed softly; a from-zero narrow-authority policy finds a different optimum that saturates its smaller envelope - the two are not the same controller in different units.
Applies when
- proposing to move a deployment derating into a training constant
- a scaled-down contract policy hugs the action clip
- Isaac-green / cross-sim-zero results on a re-scaled lineage
“预注册 a) 证伪——Isaac 全绿(零摔/reward 117)但 MuJoCo --delay 2 八档 checkpoint 扫描全数 0/3、iter6500 20-seed 0/20 … 「s1c@0.8 = 0.4 训练孪生」的代数等价不成立: 部署端打折的策略活在 0.5 的内模里,训练起点收权限学出的是贴 clip 的另一套步态,对 plant 差异零余量。”
train/OMNI_V0_SPEC.md § 3. S1.6 判决(2026-08-07 验收)