{
 "credit": "Taylor (gtmcknight), github.com/gtmcknight/amm-sim@4812a6382311624e1506aca131ca35c378f930d9 research/solved: the solved policy, Gaussian filter, qm tables, exact-arithmetic arb/retail tell and U1 tables that LEAN and QLX5 are built on. QLX5 (LEAN plus the drift-tolerant split test) is code_zb64.QLX5 of https://api.imd.fun/artifacts/1ef80bfb6d6791b877b1f4048a5e63cf2ae54153e6e8763ae3f8ab13132dfb15; the knob is entry n=8 of the c-map result (sha256 f921f2133d55d8ce81b5d809494003e898fc2a72d1fe47cdc34606ff269d676e); block driver: driver_zb64 of https://api.imd.fun/artifacts/764c9461a9f3bab6e277178f4bb48553d1b97c6cca33bed016f7c9bfd611cc42; kit: identity-md-launches/goal-3178-build-shared-evaluation-kit@45e7563d1a29ffca9d608fdf60c90bf2624d2d01. This item only moves one constant of that source; all credit for the policy is Taylor's.",
 "n": 8,
 "id": "pred_var",
 "family": "table levels",
 "find": "(fb,fa)=fees(w,rr,ph,syn,v);",
 "lo": "(fb,fa)=fees(w,rr,ph,syn,v+446512500000);",
 "hi": "(fb,fa)=fees(w,rr,ph,syn,v+893025000000);",
 "seeds": {
  "first": 10204024,
  "last": 10204151,
  "n": 128,
  "unused": []
 },
 "arms": {
  "QLX5": {
   "mean_edge": 535.139862798844,
   "peak_gas": 207233,
   "out_of_gas": 0,
   "runtime_bytes": 22440,
   "source_sha256": "e183e5bc3fb794b7067703b10a7da9920267aa5f4caa968e16ca101b02338f40",
   "source_chars": 47230,
   "runtime_sha256": "903a0cf860910022916e7c16aa1ffcfb94b0db9e25074665947d786615a44ae8",
   "reverts_halts_errors": 0,
   "afterSwap_calls": 1396536,
   "mean_gas": 136608.7,
   "engine_seconds": 34.017
  },
  "LO": {
   "mean_edge": 534.8999294426309,
   "peak_gas": 205921,
   "out_of_gas": 0,
   "runtime_bytes": 22447,
   "source_sha256": "2fcc1d88a70fbbdbcae525219914d667ff695e45f966ba0150b66b439c5ed186",
   "source_chars": 47243,
   "runtime_sha256": "42166ea237c7ab155fbda280adae1c850555dac7b599cca00bc58cc9e752eed3",
   "reverts_halts_errors": 0,
   "afterSwap_calls": 1370051,
   "mean_gas": 137697.4,
   "engine_seconds": 34.684
  },
  "HI": {
   "mean_edge": 534.5569563559304,
   "peak_gas": 207492,
   "out_of_gas": 0,
   "runtime_bytes": 22447,
   "source_sha256": "a5ce466a646f2498e9dbf739430f0a86e5e59a5d8252f8994ef6cfe3b803ade5",
   "source_chars": 47243,
   "runtime_sha256": "30b20e93996e6cac5afb689e9ba3f9128fdecfda8f074cfed72d09340b676573",
   "reverts_halts_errors": 0,
   "afterSwap_calls": 1354266,
   "mean_gas": 137918.5,
   "engine_seconds": 33.333
  }
 },
 "diffs": {
  "LO-QLX5": {
   "n": 128,
   "mean_diff": -0.23993335621308765,
   "se": 0.03929320255140311,
   "sum_d": -30.71146959527522,
   "sum_d2": 32.46725091844692,
   "seeds_positive": 34,
   "min": -1.1706945262974386,
   "max": 1.2757620622456898
  },
  "HI-QLX5": {
   "n": 128,
   "mean_diff": -0.5829064429135431,
   "se": 0.05811271005553819,
   "sum_d": -74.61202469293352,
   "sum_d2": 98.38975732223989,
   "seeds_positive": 22,
   "min": -2.6998573329130977,
   "max": 1.5781432739403272
  },
  "HI-LO": {
   "n": 128,
   "mean_diff": -0.34297308670045545,
   "se": 0.031688539643463014,
   "sum_d": -43.9005550976583,
   "sum_d2": 31.38039147292476,
   "seeds_positive": 20,
   "min": -1.960109307730761,
   "max": 0.7077627759109646
  }
 },
 "pass_peak": {
  "LO": true,
  "HI": true
 },
 "reading": "Adding filter diffusion to the variance handed to fees() lowers engine edge monotonically on these 128 fresh seeds: LO - QLX5 = -0.2399 (se 0.0393, 34 of 128 seeds positive) and HI - QLX5 = -0.5829 (se 0.0581, 22 of 128 positive), about 6 and 10 standard errors below zero, so QLX5's posterior v should stay and neither variant is a candidate.",
 "method": {
  "build": "QLX5 decoded from code_zb64.QLX5 (47230 characters, source sha256 checked equal to e183e5bc3fb794b7067703b10a7da9920267aa5f4caa968e16ca101b02338f40); find asserted to occur exactly once; LO and HI are that source with find replaced by lo and by hi, nothing else changed; their source sha256, characters (47243) and runtime bytes (22447) equal the c-map entry's.",
  "compile": "driver compile_arm: kit SolidityCompiler, pinned solc 0.8.24 (binary sha256 fb03a29a517452b9f12bcf459ef37d0a543765bb3bbc911e70a87d6a37c30d5f), optimizer runs 200, viaIR, evmVersion paris; QLX5 runtime sha256 equals the one A/1ef80bfb reports.",
  "engine": "kit setup_fast.sh --platform linux-x86_64 (exit 0, 16 pinned downloads): amm-challenge@6b6921f8d463314aa7c96234e1be25a25245f704, amm_sim_rs 0.1.0.post1, native sha256 ff88b6353c2aad7342f0a26194f7347e78b9c3b38d683a7449006d423978cd14",
  "driver": "driver_zb64 of A/764c9461 decoded (19802 bytes, sha256 d3f8646b541a76921a47e9d03247c8c9dfab594076a7c5dca79c36b6c048a282, equal to its driver_sha256), unedited: one command, --block 32 --lease 10204024 10204151, arms QLX5 then LO then HI inside each block, each arm alone against the 30 bps VanillaStrategy, workers 8; exit 0; configs checked equal to the kit's on all 128 seeds.",
  "blocks_complete": 4,
  "blocks": [
   {
    "seeds": [
     10204024,
     10204055
    ],
    "LO-QLX5": -0.27804,
    "HI-QLX5": -0.685396
   },
   {
    "seeds": [
     10204056,
     10204087
    ],
    "LO-QLX5": -0.241425,
    "HI-QLX5": -0.509342
   },
   {
    "seeds": [
     10204088,
     10204119
    ],
    "LO-QLX5": -0.156329,
    "HI-QLX5": -0.463253
   },
   {
    "seeds": [
     10204120,
     10204151
    ],
    "LO-QLX5": -0.283939,
    "HI-QLX5": -0.673635
   }
  ],
  "engine_seconds_total": 102.034,
  "ledger": {
   "seeds_complete": 128,
   "burnt": [],
   "failures": [],
   "partial": [],
   "outside_lease": 0
  },
  "stats": "edge = engine edges[\"submission\"], unrounded; diffs are per-seed paired differences; se = sd(n-1)/sqrt(n); peak_gas = largest successful Submission afterSwap gas over the 128 seeds; reverts_halts_errors counts oog, revert, other_halt, tx_error and bad_return over every callback of both pools."
 },
 "not_done": [
  "No gate (P, Q, T, F, G, E) was run, no replica run was made and nothing was submitted to the leaderboard; the arms were paired with QLX5 only, not with qt1.",
  "Only the two values of the map were tried (v + half and v + one diffusion step); values below the posterior v (a negative add or a scale under 1) were not tested, so this does not show that v is the optimum, only that moving up from it loses edge.",
  "Peaks and the zero out-of-gas count come from these 128 seeds only.",
  "No seed outside 10204024..10204151 was simulated; every seed was run once per arm, none burnt, failed, partial or unused.",
  "Per-seed edges are not in this file (kept in the local ledger only, which is not delivered); sum_d and sum_d2 are given so the differences can be pooled."
 ]
}