{"status":"completed: 200 CMA generations from START restart; frozen best-monitor weights (gen 180, es_mean) valid=True; real paired vs IB +0.128 se 0.017 on 512 fresh seeds; replica holdout +0.103 se 0.005","blocked":null,"item":"m-s22-p32","batch":"ib-r1","start":{"id":"afd5183e9ff3235334e041c00b0cd19079599a2c6464db95f8c1586712105557","url":"https://api.imd.fun/artifacts/afd5183e9ff3235334e041c00b0cd19079599a2c6464db95f8c1586712105557","code_sha256":"301941829a44bb60977bb28181a5e4bfa6544f3696763f987054789b2611d686","note":"START is the IB file itself, so START weights = IB weights"},"hashes":{"start_json_sha256_equals_id":true,"code_sha256_verified":true,"port_sha256_verified":true,"rep_json_sha256_equals_id":true,"ha_json_sha256_equals_id":true,"port_json_sha256_equals_id":true,"artifact_urls":{"ha":"https://api.imd.fun/artifacts/2801efc8190e86318dd922d4a7179f63684d77a226047fc2f3c78232963eb690","t":"https://api.imd.fun/artifacts/6ac5e013b7b1fb4c5e7bdcdfec0aba28b16824e18ccf39500f258533879faffe","rep":"https://api.imd.fun/artifacts/e8a7a2af9039c24b8bb2d3b5fa1c9113606cbf5a480ac07d02ca26750f1752ea","port":"https://api.imd.fun/artifacts/d845a4b94eab87f23956fd1939c0135d955df212d1f3037f2a761c178cef7a0e"},"repo_commit":"6b6921f8d463314aa7c96234e1be25a25245f704","wheel_sha256":"aedb68280aca34c110968f00148387c95885207913fb2cf19a1ab4c5cfd6887a"},"code_zb64":"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","code_sha256":"ccce101be7196062a271a49b5647e4eff41cdf8ed477dc0ad153b0b7b10cb994","contract_name":"inputs_b","port_zb64":"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","port_sha256":"f63227c5e0f63633efd132d6fb402ede665e5c9cb24481911471e8dc20a3362d","port_notes":"ha_port.py = RS2 port + T toxicity anchor (S[16] mean, S[17] count; cap 674, k_lin 0.712696014807650816, k_quad 1.295240753826310656) + 19-8-2 head whose weights are TH[HEAD0:HEAD0+178] (HEAD0=len(RS2 TH)+3). Same strat_init/strat_after API; plug into the replica STRATEGY block via import. Weights zero => exactly T; HA weights with new columns zero => HA.","params":{"W1":[[0.856687,1.530214,-0.918179,3.154883,-2.814956,-0.912827,-0.943631,1.469639,-4.991061,0.000158,1.693662,-0.428254,0.273345,1.271718,0.306263,-1.335815,0.318898,0.009955,0.467764],[0.448749,1.339697,2.616729,2.762034,0.457537,-0.875158,-0.112055,-0.111286,-0.324037,-0.010048,-0.052975,-0.653065,0.213663,-1.275637,-0.826425,1.456854,0.527137,-0.636551,0.257454],[1.299314,2.985459,0.293377,0.464858,0.634261,1.64482,-1.604576,1.189232,0.740338,-0.677701,0.158426,-3.071541,-0.862417,-0.187092,0.550869,0.992269,-0.72645,-0.90781,-0.817792],[-1.020102,-0.625389,1.387789,-0.300842,-0.504001,1.228704,-0.988518,1.758471,-1.998577,-0.34062,0.427317,-2.342075,-1.779936,-0.30883,0.322273,0.492864,1.842629,0.410684,2.176646],[0.152401,0.864602,2.342864,-0.115044,-2.308975,-0.051329,1.244358,1.163663,-0.094327,0.123686,-0.000559,-1.786478,-0.143573,-0.454986,-0.013734,-1.424883,2.025029,-0.484498,-0.12752],[2.546728,-1.134585,1.899408,-1.016335,1.580895,1.359446,0.043393,-0.126944,-0.643144,0.088783,-0.283905,3.157029,1.143716,-0.488838,-1.172663,0.35281,-0.031914,1.344926,0.148757],[-1.229171,1.430531,-0.008204,-2.966867,1.52344,-0.127698,-0.517299,0.077092,0.924555,-0.169457,0.950602,0.333615,-0.664494,0.40114,0.425482,0.242172,0.095775,0.74543,-0.239084],[-2.162442,-1.654734,2.772118,2.752616,-0.369154,-2.497215,-1.196389,2.851819,1.218994,-1.265298,2.864027,-2.129389,0.341858,0.421407,-0.088376,2.349241,0.695387,-0.173671,-0.350267]],"b1":[1.853683,-2.129946,-3.281985,-1.801885,-0.77521,-1.845071,-2.832558,-0.973152],"W2":[[0.52406,-1.536958,0.341368,-0.392099,-2.913296,-0.640747,-2.095228,0.741303],[0.204695,-1.596304,-1.331356,0.383072,2.043468,-1.031753,0.668939,0.458792]],"b2":[2.007957,1.972328],"hidden_units":8,"input_count":19,"layout":"flat vector = W1 (8x19 row-major), b1(8), W2 (2x8 row-major: bid row then ask row), b2(2)","weight_fixed_point_scale":1000000,"feature_fixed_point_scale":1000000,"output_unit_fee":0.0001,"feature_order":["sigmaHat","lambdaHat","sizeHat","toxEma","actEma","direction","(spot-staleHat)/staleHat","signed_trade_ratio","arb_probability","first_in_step","uncertainty","min(silent,8)/8","anchor_mean","raw_amountY","step_trade_count","prev_bid_fee","prev_ask_fee","steps_since_arb_like_print","router_implied_vanilla_spot"],"feature_multipliers":[1000,2,500,500,500,1,500,500,2,2,1000000,1,500,0.05,0.5,333.333333,333.333333,0.125,500],"feature_offsets":[-1,-1,-1,-1,-1,-1,0,0,-1,-1,-1,0,-1,-1,-0.5,-1,-1,-1,0],"feature_clip":[-4,4],"hidden_relu_clip":[0,4],"toxicity_anchor":{"count_cap":674,"linear_k_wad":712696014807650816,"quadratic_k_wad":1295240753826310656}},"new_input_formulas":{"raw_amountY (x13)":"trade.amountY/20 - 1 (raw, uncapped Y amount of the print; fixed: amountY/20/1e12 - 1e6)","step_trade_count (x14)":"(stepTradeCount - 1)*0.5 where stepTradeCount is the in-step count after this print, capped 64","prev_bid_fee (x15) / prev_ask_fee (x16)":"fee_before_this_callback/0.003 - 1 (slots[0]/slots[1] read at callback entry)","steps_since_arb_like_print (x17)":"min(timestamp - lastArbTs, 16)*0.125 - 1; a print is arb-like when it is first in its step and the posterior arbProbability >= 0.5; lastArbTs updated on the current print before the feature is computed, 0 before any","router_implied_vanilla_spot (x18)":"on each later-in-step (retail) print, v = isBuy ? spot*(1-feeUsed)/(1-0.003) : spot*(1-0.003)/(1-feeUsed) (router equalizes fee-adjusted marginal prices when the order splits); stored and held; feature = (v - spot)/spot*500, 0 before the first retail print","storage":"lastArbTs (64 bits at <<80) and routerSpot (bits >=144) are packed into slots[16] with the anchor mean (low 64) and count (uint16 at <<64); no new storage slot","clip":"all 19 inputs clipped to [-4,4] like HA","dropped_unobservable":"The true Vanilla spot and true fair price are not callback-observable, so only the router-implied estimate is used; a print is only labelled arb-like via the posterior probability, never via simulator truth. Nothing else was dropped: all requested inputs are observable from TradeInfo and own storage."},"gates":{"A":{"sequences":50,"seeds":[310000,310049],"swap_callbacks":100000,"callbacks_incl_init":100050,"max_fee_diff_bps":6.64202309125228e-05,"fees_over_0.01bps":0,"threshold_bps":0.01,"passed":true,"what":"START port (START weights) vs START Solidity in EVM (adapted differential_test_source)"},"B":{"n":512,"mean_diff":0.5863817609921741,"se":0.03113175959364559,"abs_error_vs_real_0.592":0.005618239007825876,"seeds":[7132768,7133279],"real_reference":0.592,"passed":true,"what":"replica START minus HA (HA = HA weights, 6 new columns 0)"},"C":{"sequences":50,"seeds":[310000,310049],"swap_callbacks":100000,"callbacks_incl_init":100050,"max_fee_diff_bps":6.968931338252604e-05,"fees_over_0.01bps":0,"threshold_bps":0.1,"passed":true,"what":"frozen Solidity vs START port with frozen weights"}},"speed":{"markets_per_s":786.0641123009802,"processes":6,"scope":"512 x 10000-step replica markets with START port, warmed Numba, persistent fork pool (seeds 7132768..7133279)","search_throughput_markets_per_s":840.2,"search_throughput_note":"approx: ~19.5 s per 32x512-market generation incl. monitor overhead"},"search":{"method":"CMA-ES (pycma), all 178 head parameters","es_seed":20261102,"population":32,"sigma0":0.022,"CMA_stds":[1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.600727,1.856486,1.856486,1.856486,1.856486,1.856486,1.856486,1.856486,1.856486,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.272622,1.185932,1.185932],"CMA_stds_rule":"per-group RMS of HA weights (W1 incl. new columns = rms(HA W1), b1, W2, b2), as IB search.CMA_stds describes","start":"START search.restart.es_mean_at_last_gen","generations":200,"train_seeds":"502000000 + 512*g .. +511, g=0..199 (used 502000000..502102399)","monitor_seeds":[30000000,30002047],"monitor_every":10,"curve_columns":["gen","monitor_gain_vs_START_es_mean","se","monitor_gain_vs_START_gen_best","train_mean_edge_gen_best","train_mean_edge_population","sigma","best_monitor_so_far"],"curve":[[10,0.0191,0.007,-0.0007,528.3254,528.2905,0.01852,0.0191],[20,0.0206,0.0071,0.0198,525.2423,525.1954,0.01788,0.0206],[30,0.0293,0.0072,0.0088,529.9322,529.8923,0.01811,0.0293],[40,0.0328,0.0073,0.0295,525.2689,525.2311,0.01846,0.0328],[50,0.0433,0.008,0.0398,528.2584,528.2198,0.01834,0.0433],[60,0.0584,0.0092,0.0479,530.1277,530.0987,0.01865,0.0584],[70,0.055,0.009,0.0434,528.6379,528.6061,0.01877,0.0584],[80,0.0528,0.0089,0.0332,523.2455,523.1974,0.0191,0.0584],[90,0.0701,0.0091,0.045,531.4834,531.4388,0.01856,0.0701],[100,0.0566,0.0087,0.0509,528.6119,528.5798,0.01776,0.0701],[110,0.0747,0.009,0.0797,531.4123,531.3786,0.01727,0.0797],[120,0.0812,0.0094,0.0682,528.6822,528.6572,0.01724,0.0812],[130,0.0961,0.0091,0.0829,526.4904,526.4435,0.01717,0.0961],[140,0.0967,0.0092,0.0726,535.5992,535.5753,0.01685,0.0967],[150,0.0984,0.0093,0.0867,529.2803,529.2428,0.01713,0.0984],[160,0.1009,0.0093,0.0469,526.7769,526.7373,0.01694,0.1009],[170,0.105,0.0095,0.095,537.5263,537.4926,0.01682,0.105],[180,0.113,0.0096,0.0894,532.9982,532.9625,0.01595,0.113],[190,0.1033,0.009,0.1051,528.3708,528.3474,0.01498,0.113],[200,0.1057,0.0096,0.0931,525.3087,525.2758,0.01444,0.113]],"monitor_START_mean_edge":528.2419735073306,"selection":"best monitor gain vs START among ES mean and generation-best proposal every 10 gens; START itself was the incumbent (gain 0)","frozen":{"gain_vs_START_monitor":0.11295503402189427,"se":0.00964660421517996,"gen":180,"src":"es_mean"},"weights_rounded":1e-06,"restart":{"frozen_weights":"= params","es_mean_at_last_gen":[0.859196,1.554372,-0.943415,3.100273,-2.848667,-0.912205,-0.879199,1.505538,-5.023785,0.007195,1.687402,-0.336493,0.267117,1.285556,0.286724,-1.337439,0.32602,-0.002918,0.47201,0.436613,1.34545,2.645119,2.837886,0.449984,-0.85846,-0.094665,-0.079845,-0.300766,-0.030607,-0.071333,-0.679239,0.2532,-1.262179,-0.768725,1.461243,0.520278,-0.601745,0.27454,1.23132,2.95493,0.253319,0.491165,0.674425,1.605465,-1.63575,1.161579,0.739156,-0.694619,0.154806,-3.086206,-0.829716,-0.174503,0.541664,0.929064,-0.765366,-0.896286,-0.819866,-0.966347,-0.596232,1.397639,-0.221536,-0.538732,1.191275,-0.97257,1.742626,-2.002161,-0.342835,0.391496,-2.289697,-1.83399,-0.323203,0.315986,0.482548,1.778873,0.395142,2.107347,0.135514,0.812712,2.371595,-0.091739,-2.39423,-0.061889,1.243197,1.164346,-0.06854,0.103786,0.007535,-1.75368,-0.143149,-0.433168,0.002947,-1.424765,2.058209,-0.500326,-0.144376,2.562549,-1.144779,1.977541,-1.050392,1.594475,1.339392,0.075256,-0.086803,-0.640692,0.12626,-0.28402,3.128347,1.178648,-0.468775,-1.138224,0.36892,-0.059246,1.332332,0.102999,-1.244686,1.465248,-0.0174,-2.935497,1.497118,-0.129186,-0.495195,0.022872,0.898909,-0.158609,0.95964,0.289457,-0.656225,0.432324,0.453774,0.35359,0.105894,0.755401,-0.227883,-2.172968,-1.626419,2.822767,2.800568,-0.402061,-2.483649,-1.227449,2.865762,1.241177,-1.259492,2.869437,-2.147457,0.334012,0.415952,-0.055349,2.276162,0.698449,-0.105247,-0.391485,1.812233,-2.123135,-3.353796,-1.795589,-0.768972,-1.807155,-2.793446,-0.980205,0.525088,-1.53005,0.378053,-0.438533,-2.983023,-0.663076,-2.123797,0.778358,0.203962,-1.600829,-1.371231,0.351784,2.056246,-1.050818,0.706531,0.461646,1.98316,1.982335],"es_sigma_at_last_gen":0.014443493319630015,"last_gen":200,"next_train_seed":502102400},"still_rising":"best monitor rose from +0.019 (gen 10) to +0.113 (gen 180); gens 190-200 did not beat it; sigma decaying 0.0185 -> 0.0144"},"replica_validation":{"holdout_seeds":[41000000,41008191],"evaluated_once":true,"frozen_vs_start":{"n":8192,"mean_diff":0.10264958576904631,"se":0.004778748064578869},"frozen_vs_IB":{"n":8192,"mean_diff":0.10264958576904631,"se":0.004778748064578869},"note":"START weights = IB weights, so both comparisons are the same pair","retail_edge_delta":-0.08093533908222592,"arb_edge_delta":0.18358492485127176,"means":{"frozen":527.1458757793446,"start":527.0432261935755}},"real":{"seeds":[8001024,8001535],"chunk":64,"engine":"amm_sim_rs.run_batch(bytecode, VanillaStrategy, configs, 8), IB and frozen each alone; configs = MatchRunner._build_configs body (BASELINE_SETTINGS/BASELINE_VARIANCE) with seed i -> seed","paired_vs_IB":{"n":512,"mean_diff":0.1276070276059561,"se":0.01745856910853318,"worst5":[[8001377,-1.4360655664694377],[8001282,-1.3093711445025065],[8001081,-1.2243226940094019],[8001432,-1.1497366237160804],[8001420,-1.0890034193811289]]},"retail_volume_y_delta":-83.96416555402335,"arb_volume_y_delta":-77.22120505640832,"pnl_delta":0.13239645863599492,"retail_arb_note":"LightweightSimResult exposes edges, pnl, retail_volume_y, arb_volume_y, average_fees but no retail/arb edge split; real deltas are volume deltas (frozen minus IB). Replica retail/arb edge deltas are in replica_validation.","mean_edge":{"IB":527.7775214607084,"frozen":527.9051284883144},"probe":"one 8-seed probe of START on 8000000..8000007 to check the harness; not reused"},"gas":{"afterSwap_peak":234896,"oog":0,"other_failures":0,"IB_afterSwap_peak":234884,"IB_oog":0,"gas_stats":{"frozen":[{"role":"Baseline","callback":"afterInitialize","calls":512,"ok":512,"max_ok_gas":21796,"sum_ok_gas":11159552,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Baseline","callback":"afterSwap","calls":5492042,"ok":5492042,"max_ok_gas":22592,"sum_ok_gas":123915795328,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Baseline","callback":"getName","calls":512,"ok":512,"max_ok_gas":21526,"sum_ok_gas":11021312,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Submission","callback":"afterInitialize","calls":512,"ok":512,"max_ok_gas":243227,"sum_ok_gas":124532224,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Submission","callback":"afterSwap","calls":5506925,"ok":5506925,"max_ok_gas":234896,"sum_ok_gas":744773171354,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Submission","callback":"getName","calls":512,"ok":512,"max_ok_gas":21735,"sum_ok_gas":11128320,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0}],"IB":[{"role":"Baseline","callback":"afterInitialize","calls":512,"ok":512,"max_ok_gas":21796,"sum_ok_gas":11159552,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Baseline","callback":"afterSwap","calls":5482688,"ok":5482688,"max_ok_gas":22592,"sum_ok_gas":123704734420,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Baseline","callback":"getName","calls":512,"ok":512,"max_ok_gas":21526,"sum_ok_gas":11021312,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Submission","callback":"afterInitialize","calls":512,"ok":512,"max_ok_gas":243227,"sum_ok_gas":124532224,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Submission","callback":"afterSwap","calls":5507369,"ok":5507369,"max_ok_gas":234884,"sum_ok_gas":744850984774,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0},{"role":"Submission","callback":"getName","calls":512,"ok":512,"max_ok_gas":21735,"sum_ok_gas":11128320,"oog":0,"revert":0,"other_halt":0,"tx_error":0,"bad_return":0}]}},"valid":true,"commands":["git clone github.com/benedictbrady/amm-challenge @6b6921f8 under test/scratch/amm-challenge; sh scratch/setup.sh (verbatim, succeeded, StarterStrategy validated)","python diff.py <sol> <weights> <seq0> <seq1> (adapted differential_test_source with ha_port; 5 shards x 10 sequences, seeds 310000..310049)","python gateb.py (gate B + speed)","python search.py <wall_s> [gmax] (10 segments < 8 min, checkpoint ckpt.pkl)","python freeze.py; amm-match validate frozen.sol (ok); python holdout.py; python real.py out 8001024 512 <sol> per arm"],"notes":["All downloaded JSONs match their URL ids; START code and port decode to their stated sha256; gen_sol(START weights) reproduces START Solidity byte-for-byte, so the frozen Solidity differs from START only in the _head weight constants.","Machine has 6 cores; replica 786 markets/s, ~19.5 s per generation; 200 generations (~76 min of segments). First segment (gens 0-21) was accidentally replayed once from scratch due to a checkpoint-path slip; the replay used the same seeds and reproduced identical results, so the train seeds per generation are unique within the kept run.","Monitor gains are vs START (the IB weights) on monitor seeds 30000000..30002047, not vs T as in the IB curve.","Real engine: frozen minus IB +0.1276 (se 0.0175) on 512 fresh seeds; replica holdout predicted +0.1026.","Replica retail/arb split of the gain: retail edge slightly lower, arb loss smaller (fewer losses to arbitrage).","No leaderboard placement claimed."]}