{ "approach": "First-principles asymmetric fair-value quotes plus separate bid/ask toxicity accumulators. Each trade adds SIZE_GAIN*(amountY/reserveY)^2 and a quadratic decayed same-direction streak surcharge to its side. Pressures decay with elapsed steps, are bounded, and use eight of the 32 slots. Fair price is a smoothed fee-band-edge inference from first-in-step trades. No external calls, assembly, or history loops.", "tried": [ { "label": "A", "parameters": { "BASE": 2500, "ALPHA": 260, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "score": { "n": 1000, "mean_edge": 480.5497197727031 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 479.42046216127903 }, "paired": { "reference": "test/scratch/simulator/contracts/src/StarterStrategy.sol", "reference_sha256": "8f2cea337f6736105f58c6758da6c3ebf002951d34492bcd5b7101bef2d1e7f3", "seed_start": 0, "n": 2000, "mean_diff": 110.61604154359331, "se": 0.787410080523976, "worst": [ { "seed": 1046, "diff": -1.6016755973273007 }, { "seed": 1905, "diff": 5.348833245127707 }, { "seed": 1668, "diff": 12.427654672842493 }, { "seed": 734, "diff": 19.926585937869447 }, { "seed": 1732, "diff": 21.978242173121885 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.26610064638598124 }, { "label": "B", "parameters": { "BASE": 2500, "ALPHA": 600, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "c22645615e226ce4e28fd42895b3c98c5116fdff99e7daf5da4ea604cf70125f", "score": { "n": 1000, "mean_edge": 427.75770858115914 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 426.553961863498 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -52.86650029778101, "se": 0.4525629017322739, "worst": [ { "seed": 1270, "diff": -153.89683272862885 }, { "seed": 896, "diff": -149.1478727780197 }, { "seed": 1164, "diff": -147.54453611574297 }, { "seed": 1501, "diff": -137.96363015523661 }, { "seed": 1865, "diff": -130.36558497814224 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.5815750112774549 }, { "label": "C", "parameters": { "BASE": 3000, "ALPHA": 260, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "7f06e747579a62f2a881481363c5759ea0c5873458c039b2b7fcfc9c84abb8ec", "score": { "n": 1000, "mean_edge": 474.93928359057054 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 474.015961187373 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -5.404500973905943, "se": 0.23684810398403158, "worst": [ { "seed": 332, "diff": -106.68011535880146 }, { "seed": 14, "diff": -99.2677789948641 }, { "seed": 1933, "diff": -69.418609591821 }, { "seed": 1241, "diff": -57.43223625926953 }, { "seed": 1291, "diff": -52.87391316882247 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.48766309323644963 }, { "label": "D", "parameters": { "BASE": 2500, "ALPHA": 260, "SIZE_GAIN": 300, "STREAK_GAIN": 50, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "d3d4ba7557352a064ca9a5763883b69709e71eab81c82ee8f0b8b54b853f6a8d", "score": { "n": 1000, "mean_edge": 459.8482645926341 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 459.0722568309237 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -20.348205330355235, "se": 0.36546588046895967, "worst": [ { "seed": 634, "diff": -136.30853202949532 }, { "seed": 542, "diff": -128.7438008915591 }, { "seed": 14, "diff": -117.84642583606188 }, { "seed": 310, "diff": -116.89210086455296 }, { "seed": 1562, "diff": -110.5947888556791 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.6948160432399599 }, { "label": "E", "parameters": { "BASE": 2500, "ALPHA": 260, "SIZE_GAIN": 0, "STREAK_GAIN": 0, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "bed6efa78b24c0941c4a209cdcc1113bd6442680233a19c588f642c28a7db983", "score": { "n": 1000, "mean_edge": 481.7105637093854 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 480.57309836077235 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": 1.1526361994934078, "se": 0.18515662953363632, "worst": [ { "seed": 1562, "diff": -102.6741461221527 }, { "seed": 1724, "diff": -50.19697513678017 }, { "seed": 238, "diff": -45.465015728944365 }, { "seed": 406, "diff": -39.04968885623538 }, { "seed": 1612, "diff": -38.65474361406439 } ] }, "role": "no-toxicity ablation; not eligible as the requested toxicity implementation", "seconds_per_simulation": 0.38035826853703475 }, { "label": "F", "parameters": { "BASE": 4000, "ALPHA": 600, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "3995166e51b97669e75f304362e7c68a3f9f1b698448221bfc37f24eeeeb5667", "score": { "n": 1000, "mean_edge": 415.03346045477196 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 413.7012826953103 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -65.71917946596865, "se": 0.46401657337867086, "worst": [ { "seed": 755, "diff": -196.2765647862642 }, { "seed": 206, "diff": -195.10986369973728 }, { "seed": 1604, "diff": -176.52967093825657 }, { "seed": 1005, "diff": -148.46199271037779 }, { "seed": 944, "diff": -139.94716368467414 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.45642894724500366 }, { "label": "G", "parameters": { "BASE": 4000, "ALPHA": 260, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "afe8253dd70278b94a8f274e1aec0551de9f2e604f2b73cfda0e67d31574a62c", "score": { "n": 1000, "mean_edge": 460.29516199701976 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 459.19699404162566 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -20.223468119653397, "se": 0.3148132293497394, "worst": [ { "seed": 332, "diff": -114.6908408558516 }, { "seed": 14, "diff": -113.17297088789229 }, { "seed": 1562, "diff": -111.78825380319915 }, { "seed": 310, "diff": -97.33580601162248 }, { "seed": 1680, "diff": -97.01185880744487 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.4355601551600266 }, { "label": "H", "parameters": { "BASE": 5000, "ALPHA": 600, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "2a98241b63c1d4552e7d14d7ee5357412e4a5ebb965118f502750a06ae826e30", "score": { "n": 1000, "mean_edge": 401.9276215490016 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 400.8289242625489 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -78.59153789873007, "se": 0.4831060156700394, "worst": [ { "seed": 1270, "diff": -172.8668753260921 }, { "seed": 1431, "diff": -172.71849163858394 }, { "seed": 1226, "diff": -158.52099079722848 }, { "seed": 896, "diff": -158.42285856423746 }, { "seed": 1463, "diff": -156.60962512597706 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.6254289643990342 }, { "label": "I", "parameters": { "BASE": 2500, "ALPHA": 100, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "a8a73fcec10548f226bb9b0b2884310d3740acc567f4c2d99736814c0f6b51cc", "score": { "n": 1000, "mean_edge": 473.33655860325086 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 472.3615331859281 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -7.0589289753508355, "se": 0.38670246303893563, "worst": [ { "seed": 542, "diff": -158.77975141283173 }, { "seed": 1149, "diff": -79.82768643874414 }, { "seed": 1272, "diff": -75.99114342540219 }, { "seed": 746, "diff": -74.19101869907388 }, { "seed": 1259, "diff": -72.8445993980369 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.49598082856251857 }, { "label": "J", "parameters": { "BASE": 3000, "ALPHA": 100, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "785fb4fa539b56d64b979418f6b4bf50ddfe3dbfe106e6a6ea937dff170cfb57", "score": { "n": 1000, "mean_edge": 467.4654869594833 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 466.80401755846174 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -12.616444602817241, "se": 0.4013969138495288, "worst": [ { "seed": 542, "diff": -165.8844172228341 }, { "seed": 1259, "diff": -88.46857815335903 }, { "seed": 1149, "diff": -84.12082302570167 }, { "seed": 329, "diff": -80.62439602616735 }, { "seed": 1656, "diff": -79.05381627388442 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.6817035113014863 }, { "label": "K", "parameters": { "BASE": 3500, "ALPHA": 100, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1000 }, "code_sha256": "fb5d18d8567e5887c9105ee39db0c97ce32913d21617d758973f44e8d0fc4d45", "score": { "n": 1000, "mean_edge": 461.4085241332891 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 460.58919556629417 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -18.83126659498481, "se": 0.4065949725250449, "worst": [ { "seed": 542, "diff": -158.43309454433006 }, { "seed": 24, "diff": -116.18064588409584 }, { "seed": 1633, "diff": -96.26551913477476 }, { "seed": 1259, "diff": -89.85018541171695 }, { "seed": 1611, "diff": -88.65636976656049 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.34850991301599427 }, { "label": "L", "parameters": { "BASE": 2500, "ALPHA": 260, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 500 }, "code_sha256": "8f4f86fa6911e58125fd4c95683015cb84a615813ff2aaf9d1d64bd05e6a889e", "score": { "n": 1000, "mean_edge": 484.989044455472 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 483.77171799357325 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": 4.351255832294189, "se": 0.406591993900836, "worst": [ { "seed": 1231, "diff": -52.773656608237104 }, { "seed": 708, "diff": -51.26221527022676 }, { "seed": 791, "diff": -44.07992416736079 }, { "seed": 408, "diff": -41.84588169211361 }, { "seed": 1122, "diff": -38.88550613249265 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.9079029445830383 }, { "label": "M", "parameters": { "BASE": 2500, "ALPHA": 260, "SIZE_GAIN": 30, "STREAK_GAIN": 10, "DECAY": 1000, "SKEW": 1500 }, "code_sha256": "8970f9a5b1f342b3a36426d096dab9b2b6aa913939a81b1fbd674569e6251f79", "score": { "n": 1000, "mean_edge": 244.07967293381182 }, "training_score": { "seed_start": 0, "n": 2000, "mean_edge": 242.94747523690464 }, "paired": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 0, "n": 2000, "mean_diff": -236.47298692437437, "se": 1.261546358622159, "worst": [ { "seed": 901, "diff": -421.97536467248614 }, { "seed": 1596, "diff": -419.5617721926914 }, { "seed": 1118, "diff": -407.9057885800602 }, { "seed": 588, "diff": -400.0015744977934 }, { "seed": 1854, "diff": -396.01651854753254 } ] }, "role": "eligible configuration", "seconds_per_simulation": 0.48172412692551736 } ], "next": [ "Fit a size- and timestamp-dependent confidence for first-in-step fair-price observations; distinguish a small arbitrage correction from a retail-induced price jump before increasing the update rate.", "Measure separate buy/sell size distributions and ablate the size and streak surcharges individually on a new training block, then use a new untouched 2000-seed holdout.", "Test quadratic versus cubic pressure with exact exponential step decay and a fair-value-aligned surcharge; compare against the same frozen source with paired uncertainty." ], "code": "// SPDX-License-Identifier: MIT\npragma solidity ^0.8.24;\nimport {AMMStrategyBase} from \"./AMMStrategyBase.sol\";\nimport {TradeInfo} from \"./IAMMStrategy.sol\";\n\n/// @notice Fair-value quotes with bounded, per-side quadratic flow toxicity.\ncontract Strategy is AMMStrategyBase {\n uint256 constant BASE = 2500 * 1e12;\n uint256 constant ALPHA = 260;\n uint256 constant SIZE_GAIN = 30;\n uint256 constant STREAK_GAIN = 10 * 1e12;\n uint256 constant DECAY = 1000;\n uint256 constant SKEW = 500;\n // 0 fair price; 1/2 previous bid/ask; 3 last timestamp;\n // 4/5 buy/sell size pressure; 6 direction (1 buy, 2 sell); 7 streak.\n function afterInitialize(uint256 x, uint256 y) external override returns (uint256, uint256) {\n slots[0] = wdiv(y, x);\n slots[1] = BASE;\n slots[2] = BASE;\n slots[3] = 0;\n slots[4] = 0;\n slots[5] = 0;\n slots[6] = 0;\n slots[7] = 0;\n return (BASE, BASE);\n }\n function afterSwap(TradeInfo calldata t) external override returns (uint256 bid, uint256 ask) {\n uint256 spot = wdiv(t.reserveY, t.reserveX);\n uint256 dt = t.timestamp > slots[3] ? t.timestamp - slots[3] : 0;\n uint256 fair = slots[0];\n uint256 oldFee = t.isBuy ? slots[1] : slots[2];\n uint256 implied = t.isBuy ? wmul(spot, WAD - oldFee) : wdiv(spot, WAD - oldFee);\n // Only the first trade in a step can be the price-revealing arb.\n if (dt > 0) {\n fair = (fair * (1000 - ALPHA) + implied * ALPHA) / 1000;\n slots[0] = fair;\n }\n uint256 buyPressure = slots[4];\n uint256 sellPressure = slots[5];\n uint256 streak = slots[7];\n if (dt > 0) {\n uint256 divisor = 1000 + DECAY * (dt > 64 ? 64 : dt);\n buyPressure = buyPressure * 1000 / divisor;\n sellPressure = sellPressure * 1000 / divisor;\n streak = streak * 1000 / divisor;\n }\n uint256 ratio = wdiv(t.amountY, t.reserveY);\n if (ratio > WAD / 20) ratio = WAD / 20;\n uint256 impulse = wmul(ratio, ratio) * SIZE_GAIN;\n uint256 direction = t.isBuy ? 1 : 2;\n streak = slots[6] == direction ? streak + WAD : WAD;\n if (streak > 6 * WAD) streak = 6 * WAD;\n uint256 repeatFee = wmul(wmul(streak - WAD, streak - WAD), STREAK_GAIN);\n if (t.isBuy) buyPressure += impulse + repeatFee;\n else sellPressure += impulse + repeatFee;\n buyPressure = clamp(buyPressure, 0, 200 * BPS);\n sellPressure = clamp(sellPressure, 0, 200 * BPS);\n slots[4] = buyPressure;\n slots[5] = sellPressure;\n slots[6] = direction;\n slots[7] = streak;\n slots[3] = t.timestamp;\n uint256 divergence = wdiv(absDiff(spot, fair), fair) * SKEW / 1000;\n if (spot > fair) {\n bid = BASE + divergence + buyPressure;\n ask = BASE + sellPressure;\n ask = ask > divergence ? ask - divergence : 0;\n } else {\n ask = BASE + divergence + sellPressure;\n bid = BASE + buyPressure;\n bid = bid > divergence ? bid - divergence : 0;\n }\n bid = clampFee(bid);\n ask = clampFee(ask);\n slots[1] = bid;\n slots[2] = ask;\n }\n function getName() external pure override returns (string memory) {\n return \"Size Toxicity\";\n }\n}\n", "code_sha256": "8f4f86fa6911e58125fd4c95683015cb84a615813ff2aaf9d1d64bd05e6a889e", "score": { "n": 1000, "mean_edge": 484.989044455472 }, "paired": { "reference": "contracts/src/StarterStrategy.sol @ 6b6921f8d463314aa7c96234e1be25a25245f704", "reference_sha256": "8f2cea337f6736105f58c6758da6c3ebf002951d34492bcd5b7101bef2d1e7f3", "seed_start": 100000, "n": 2000, "mean_diff": 115.19630308108363, "se": 0.7337739067341067, "worst": [ { "seed": 101573, "diff": 59.69740816036551 }, { "seed": 100645, "diff": 60.287732310799186 }, { "seed": 100818, "diff": 60.39195821087145 }, { "seed": 100274, "diff": 62.54208611012922 }, { "seed": 100068, "diff": 62.70602642904174 } ] }, "paired_previous_best": { "reference": "Initial size-toxicity A", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 100000, "n": 2000, "mean_diff": 4.936936231888352, "se": 0.4403172299913898, "worst": [ { "seed": 101790, "diff": -136.8411722688138 }, { "seed": 100096, "diff": -64.22664481318748 }, { "seed": 101015, "diff": -47.091964648753105 }, { "seed": 100145, "diff": -46.754335605498 }, { "seed": 100930, "diff": -45.505446053988635 } ] }, "holdout_score": { "seed_start": 100000, "n": 2000, "mean_edge": 484.6976045184476 }, "seconds_per_simulation": 0.23777423722198, "validate_ok": true, "citations": [ { "url": "https://github.com/benedictbrady/amm-challenge/tree/6b6921f8d463314aa7c96234e1be25a25245f704", "use": "Pinned official simulator, strategy interface, baseline, validation, and seeded measurement protocol." }, { "url": "https://github.com/jiayaoqijia/amm-challenge-yq", "use": "Public write-up of fair-price filtering and nonlinear toxicity; no implementation copied or public score treated as local evidence." }, { "url": "https://github.com/dipankar/amm-challenge/blob/main/ANALYSIS.md", "use": "Public discussion motivating state estimation alongside trade-size signals; independent source written for this item." } ], "initial_toxicity_ablation": { "reference": "test/scratch/E.sol", "reference_sha256": "bed6efa78b24c0941c4a209cdcc1113bd6442680233a19c588f642c28a7db983", "seed_start": 0, "n": 2000, "mean_diff": -1.1526361994934078, "se": 0.18515662953363632, "worst": [ { "seed": 491, "diff": -72.51882604884531 }, { "seed": 1892, "diff": -68.66344213955927 }, { "seed": 1321, "diff": -50.157619405571154 }, { "seed": 1356, "diff": -49.044836666524816 }, { "seed": 254, "diff": -47.23636409537312 } ] }, "timing": { "final_candidate_seconds": 475.54847444396, "starter_seconds": 245.4719830059912, "final_2000_pair_seconds": 721.0204574499512, "pilot_n": 20, "pilot_candidate_seconds_per_simulation": 0.2701678659475874, "pilot_reference_seconds_per_simulation": 0.259942794899689, "predicted_2000_pair_seconds": 1060.221321694553, "pair_budget_seconds": 1500 }, "selection_rule": "Highest mean on training seeds 0..1999 among nonzero size/streak implementations; promotion requires fresh paired mean_diff > 2*se.", "selected_configuration": "L", "promotion_test": { "challenger": "L", "vs_initial": { "reference": "test/scratch/A.sol", "reference_sha256": "301844e4e51ecd53db5e0cff90a669a5ad3ed65d1156499615ffb43befdfd32e", "seed_start": 100000, "n": 2000, "mean_diff": 4.936936231888352, "se": 0.4403172299913898, "worst": [ { "seed": 101790, "diff": -136.8411722688138 }, { "seed": 100096, "diff": -64.22664481318748 }, { "seed": 101015, "diff": -47.091964648753105 }, { "seed": 100145, "diff": -46.754335605498 }, { "seed": 100930, "diff": -45.505446053988635 } ] }, "vs_starter": { "reference": "test/scratch/simulator/contracts/src/StarterStrategy.sol", "reference_sha256": "8f2cea337f6736105f58c6758da6c3ebf002951d34492bcd5b7101bef2d1e7f3", "seed_start": 100000, "n": 2000, "mean_diff": 115.19630308108363, "se": 0.7337739067341067, "worst": [ { "seed": 101573, "diff": 59.69740816036551 }, { "seed": 100645, "diff": 60.287732310799186 }, { "seed": 100818, "diff": 60.39195821087145 }, { "seed": 100274, "diff": 62.54208611012922 }, { "seed": 100068, "diff": 62.70602642904174 } ] }, "initial_vs_starter": { "reference": "test/scratch/simulator/contracts/src/StarterStrategy.sol", "reference_sha256": "8f2cea337f6736105f58c6758da6c3ebf002951d34492bcd5b7101bef2d1e7f3", "seed_start": 100000, "n": 2000, "mean_diff": 110.25936684919526, "se": 0.8591417488737265, "worst": [ { "seed": 100140, "diff": -33.154307483944194 }, { "seed": 100434, "diff": -22.533488194521055 }, { "seed": 100451, "diff": -11.522748890858281 }, { "seed": 100619, "diff": -6.151245700952245 }, { "seed": 101978, "diff": 20.057468366583237 } ] }, "promote_challenger": true }, "changes": [ "candidates/size-toxicity/Strategy.sol: complete submission contract", "candidates/size-toxicity/NOTES.md: mechanism, citations, reproduction, scoreboard, and checks", "candidates/size-toxicity/result.json: full source, hash, scores, and paired measurements", "artifacts/goal/batches/r1/items/size-toxicity/result.json: identical required JSON output" ], "checks": [ { "command": "git -C test/scratch/simulator rev-parse HEAD", "result": "6b6921f8d463314aa7c96234e1be25a25245f704" }, { "command": "maturin develop --release --manifest-path test/scratch/simulator/amm_sim_rs/Cargo.toml", "result": "Release build succeeded; unmodified pinned Rust source." }, { "command": "amm-match validate candidates/size-toxicity/Strategy.sol", "result": "PASS: static validation, optimized solc 0.8.24 compilation, deployment, initialization." }, { "command": "python test/scratch/check_callbacks.py", "result": "PASS: 2000 direct callbacks, 250000 gas limit, both directions, repeated trades, zero/tiny/large sizes, time gaps 0..100; fee bounds held." }, { "command": "python test/scratch/check_mechanism.py", "result": "PASS: quadratic size scaling, side separation, repeat escalation, pressure decay. Initial scratch check_mechanism.py:18 assertion comparing final fees across a gap failed because fair value also updates; repaired to compare stored pressure." }, { "command": "amm-match run candidates/size-toxicity/Strategy.sol --simulations 20 (initial A source)", "result": "PASS: 516.59, matching custom harness 516.5867031359468." }, { "command": "python test/scratch/bench.py candidates/size-toxicity/Strategy.sol --reference test/scratch/simulator/contracts/src/StarterStrategy.sol --n 20", "result": "Frozen-source timing pilot completed; forecast was below the 25-minute pair allowance, so the full fresh n=2000 was retained." }, { "command": "python test/scratch/bench.py test/scratch/A.sol --reference test/scratch/simulator/contracts/src/StarterStrategy.sol --n 2000", "result": "Completed paired training seeds 0..1999." }, { "command": "python test/scratch/variants.py; python test/scratch/variants2.py; python test/scratch/variants3.py; python test/scratch/variants4.py", "result": "Completed 2000 seeds per variant, exclusively 0..1999; paired summaries recorded in tried." }, { "command": "python test/scratch/holdout.py L", "result": "Completed 2000 fresh seeds 100000..101999 per measured strategy, paired against initial A and pinned StarterStrategy; no holdout-based coefficient tuning." }, { "command": "python test/scratch/audit.py", "result": "PASS: independent paired arithmetic, JSON parsing, source hash consistency, identical result copies, 64000-byte limit, and allowed-path audit" } ], "limitations": [ "Only the bounded round-1 size-toxicity item was evaluated. No overall network best source was present; the local previous best is identified in paired_previous_best. Public leaderboard results were not reproduced or submitted.", "The no-toxicity ablation E is reported transparently but does not implement the required size/streak mechanism and is excluded from delivery eligibility.", "Training wall times include CPU contention from parallel independent measurement jobs; seconds_per_simulation is the final holdout candidate wall time per simulation." ] }