{
  "approach": "Measure only incumbent B, with no candidate or oracle. Run the unmodified simulator at b244530ee2a6247b5f930830eeca9dd950da0e7e using load_strategy_factory and exactly one run_batch simulation per seed, default ChallengeConfig() (2000 steps), default ParameterVariance(), workers=1 and no sandbox. Read the sampled parameters already exposed by SimulationResult.regime; no hidden-state patch or changed simulator is needed. Compare unconditional per-simulation total, retail and arbitrage edge across competitor spreads 1..4 and empirical terciles of each continuous parameter.",
  "tried": [
    {
      "measurement": "parameter buckets only",
      "incumbent_code_sha256": "baef955ec10728839f7c3357016ec9eb0550e7ec689bbd5171b1cdd5c9e37273",
      "candidate": null,
      "paired_difference": null,
      "outcome": "Completed 1000 distinct seeds plus 20 verification replays; no tuning, promotion, or finalist seed use."
    }
  ],
  "next": [
    {
      "change": "Test spread-conditioned adverse-selection sizing: raise the arbitrage-loss multiplier kA for inferred competitor spread 1, while retaining tiny probability probes. Tune the multiplier on separate seeds, checking that saved arbitrage loss exceeds sacrificed retail edge.",
      "support": "Narrow-spread markets have a much larger arbitrage-loss share of gross retail edge than wide-spread markets. This is a descriptive regime comparison, not evidence that reducing size improves score.",
      "edge_at_stake": {
        "spread_1_count": 260,
        "spread_1_total_mean": 39.52276107669692,
        "spread_1_retail_mean": 60.59958368858538,
        "spread_1_arb_mean": -21.076822611870384,
        "spread_1_arb_se": 4.2979096447434735,
        "spread_1_gross_arb_loss_weighted_over_all_seeds": 5.479973879086299,
        "spread_1_arb_loss_fraction_of_retail": 0.3478047426890235,
        "spread_4_arb_loss_fraction_of_retail": 0.13102518174592412
      },
      "caveat": "Gross arbitrage loss is exposure, not an achievable gain; changing size also changes retail edge and inventory."
    },
    {
      "change": "Test regularized online estimates of retail arrival rate and mean notional in qsize, replacing fixed lam_r=0.25 and m=5.0. Infer only from permitted observations, accounting for fill censoring and informed fills; never read sampled parameters in a submission.",
      "support": "Retail edge differs materially across sampled flow regimes, while B uses fixed flow assumptions. Bucket differences justify studying calibration but do not establish its incremental value.",
      "edge_at_stake": {
        "notional_high_minus_low_total": 49.79535316190703,
        "notional_high_minus_low_retail": 55.50247719235145,
        "notional_high_minus_low_arb": -5.7071240305340005,
        "arrival_high_minus_low_total": 56.176113342923614,
        "arrival_high_minus_low_retail": 64.61873440661785,
        "arrival_high_minus_low_arb": -8.442621063619917,
        "low_notional_arb_loss": 18.509037388222154,
        "high_notional_arb_loss": 24.216161418756155
      },
      "caveat": "These are regime spreads and gross losses, not oracle values or predicted improvement. A future candidate must be measured with paired seeds and separate tuning/holdout."
    }
  ],
  "code": "\"\"\"Run with Python 3.10+, pinned simulator on PYTHONPATH, and B's .py path.\nUsage: python3 harness.py STRATEGY OUTPUT_JSONL START STOP [--verify]\nAll runs use the unmodified simulator. No extra packages required.\n\"\"\"\nimport dataclasses\nimport hashlib\nimport json\nimport pathlib\nimport sys\nimport time\nfrom orderbook_pm_challenge.config import ChallengeConfig, ParameterVariance\nfrom orderbook_pm_challenge.loader import load_strategy_factory\nfrom orderbook_pm_challenge.runner import run_batch\n\nEXPECTED = 'baef955ec10728839f7c3357016ec9eb0550e7ec689bbd5171b1cdd5c9e37273'\nFIELDS = ('competitor_spread_ticks', 'retail_arrival_rate', 'retail_mean_notional', 'jump_intensity', 'jump_sigma')\n\ndef run(factory, seed):\n    return run_batch(factory, base_config=ChallengeConfig(), variance=ParameterVariance(), n_simulations=1, seed_start=seed, workers=1).simulation_results[0]\n\ndef main():\n    path, output, start, stop, *flags = sys.argv[1:]\n    start, stop = int(start), int(stop)\n    assert 232000 <= start < stop <= 233000\n    assert hashlib.sha256(pathlib.Path(path).read_bytes()).hexdigest() == EXPECTED\n    factory = load_strategy_factory(path)\n    with open(output, 'w') as f:\n        for seed in range(start, stop):\n            t = time.perf_counter()\n            result = run(factory, seed)\n            seconds = time.perf_counter() - t\n            record = {k: getattr(result, k) for k in ('seed', 'total_edge', 'retail_edge', 'arb_edge', 'failed', 'error')}\n            record['parameters'] = {k: getattr(result.regime, k) for k in FIELDS}\n            record['seconds'] = seconds\n            if '--verify' in flags:\n                t = time.perf_counter()\n                replay = run(factory, seed)\n                record['verification_seconds'] = time.perf_counter() - t\n                record['identical_complete_result'] = dataclasses.asdict(result) == dataclasses.asdict(replay)\n                assert record['identical_complete_result'], seed\n                record['baseline_edges'] = [replay.total_edge, replay.retail_edge, replay.arb_edge]\n            f.write(json.dumps(record) + '\\n')\n            f.flush()\n            print(seed, round(seconds, 4), result.failed, flush=True)\n\n\n\n\"\"\"Usage: python3 summarize.py pilot.jsonl part0.jsonl ... part7.jsonl\nTerciles are empirical ranks, sorted by (parameter, seed). For n=1000,\nlow/middle/high contain 334/333/333 simulations. All SEs are sample SD/sqrt(n).\n\"\"\"\nimport hashlib\nimport json\nimport math\nimport pathlib\nimport statistics\nimport sys\n\nFIELDS = ('competitor_spread_ticks', 'retail_arrival_rate', 'retail_mean_notional', 'jump_intensity', 'jump_sigma')\n\ndef summary(rows):\n    n = len(rows)\n    out = {'count': n}\n    for src, target, se in [('total_edge', 'mean', 'se'), ('retail_edge', 'retail_mean', 'retail_se'), ('arb_edge', 'arb_mean', 'arb_se')]:\n        values = [r[src] for r in rows]\n        out[target] = statistics.mean(values) if n else None\n        out[se] = statistics.stdev(values) / math.sqrt(n) if n > 1 else None\n    return out\n\ndef aggregate(paths):\n    rows = sorted([json.loads(line) for path in paths for line in pathlib.Path(path).read_text().splitlines()], key=lambda r: r['seed'])\n    seeds = [r['seed'] for r in rows]\n    assert seeds == list(range(232000, 232000 + len(rows)))\n    assert len(rows) in (500, 1000)\n    assert all(232000 <= s < 233000 for s in seeds)\n    assert all(math.isclose(r['total_edge'], r['retail_edge'] + r['arb_edge'], rel_tol=1e-10, abs_tol=1e-8) for r in rows)\n    findings = []\n    # Membership uses all attempted seeds; failed runs, if any, are explicitly\n    # excluded from means, following BatchResult.mean_edge.\n    for field in FIELDS:\n        ordered = sorted(rows, key=lambda r: (r['parameters'][field], r['seed']))\n        if field == 'competitor_spread_ticks':\n            groups = [(str(v), [r for r in rows if r['parameters'][field] == v]) for v in range(1, 5)]\n        else:\n            groups = [(label, [r for i, r in enumerate(ordered) if (3*i)//len(ordered) == k]) for k, label in enumerate(('low', 'middle', 'high'))]\n        for label, group in groups:\n            stats = summary([r for r in group if not r['failed']])\n            findings.append({'name': field + ':' + label, 'parameter': field, 'bucket': label,\n                             'sample_min': min(r['parameters'][field] for r in group),\n                             'sample_max': max(r['parameters'][field] for r in group),\n                             'attempted': len(group), 'failures': sum(r['failed'] for r in group), **stats})\n    clean = [r for r in rows if not r['failed']]\n    contrasts = []\n    for field in FIELDS:\n        groups = [r for r in findings if r['parameter'] == field]\n        lo, hi = groups[0], groups[-1]\n        contrasts.append({'parameter': field, 'comparison': hi['bucket'] + ' minus ' + lo['bucket'],\n                          'mean_diff': hi['mean'] - lo['mean'], 'se': math.hypot(hi['se'], lo['se']),\n                          'retail_diff': hi['retail_mean'] - lo['retail_mean'],\n                          'retail_se': math.hypot(hi['retail_se'], lo['retail_se']),\n                          'arb_diff': hi['arb_mean'] - lo['arb_mean'],\n                          'arb_se': math.hypot(hi['arb_se'], lo['arb_se'])})\n    pilot = [r for r in rows if 'identical_complete_result' in r]\n    assert len(pilot) == 20 and all(r['identical_complete_result'] for r in pilot)\n    return {'n': len(rows), 'incumbent': summary(clean), 'findings': findings, 'contrasts': contrasts,\n            'failures': [{'seed': r['seed'], 'error': r['error']} for r in rows if r['failed']],\n            'seconds_per_simulation': statistics.mean(r['seconds'] for r in rows),\n            'pilot_seconds_per_simulation': statistics.mean(r['seconds'] for r in pilot),\n            'pilot_verification': [{'seed': r['seed'], 'baseline': r['baseline_edges'],\n                                    'captured': [r['total_edge'], r['retail_edge'], r['arb_edge']],\n                                    'identical_complete_result': r['identical_complete_result']} for r in pilot],\n            'raw_records_sha256': hashlib.sha256(json.dumps(rows, sort_keys=True, separators=(',', ':')).encode()).hexdigest()}\n\n\nif __name__ == '__main__':\n    if len(sys.argv) > 1 and sys.argv[1] == '--summarize':\n        print(json.dumps(aggregate(sys.argv[2:]), indent=2))\n    else:\n        main()\n",
  "code_sha256": "537f967b2a26d2aa3c7ceebb69bf935f9888ca4dfa85bbcb7e93af239f857665",
  "incumbent_code_sha256": "baef955ec10728839f7c3357016ec9eb0550e7ec689bbd5171b1cdd5c9e37273",
  "tuning": null,
  "seeds": {
    "start": 232000,
    "end_inclusive": 232999,
    "verification_replayed_start": 232000,
    "verification_replayed_end_inclusive": 232019,
    "forbidden_finalist_seeds_run": false
  },
  "n": 1000,
  "complete": true,
  "incumbent": {
    "count": 1000,
    "mean": 98.308080281751,
    "se": 5.30736292666585,
    "retail_mean": 121.3989136260664,
    "retail_se": 5.5990324008476104,
    "arb_mean": -23.090833344324,
    "arb_se": 1.5751318631457436
  },
  "findings": [
    {
      "name": "competitor_spread_ticks:1",
      "parameter": "competitor_spread_ticks",
      "bucket": "1",
      "sample_min": 1,
      "sample_max": 1,
      "attempted": 260,
      "failures": 0,
      "count": 260,
      "mean": 39.52276107669692,
      "se": 7.008355130308519,
      "retail_mean": 60.59958368858538,
      "retail_se": 7.815288276984682,
      "arb_mean": -21.076822611870384,
      "arb_se": 4.2979096447434735
    },
    {
      "name": "competitor_spread_ticks:2",
      "parameter": "competitor_spread_ticks",
      "bucket": "2",
      "sample_min": 2,
      "sample_max": 2,
      "attempted": 250,
      "failures": 0,
      "count": 250,
      "mean": 71.7659088850952,
      "se": 7.617734657988405,
      "retail_mean": 93.2407364520572,
      "retail_se": 8.414715776487153,
      "arb_mean": -21.4748275669524,
      "arb_se": 2.4431444345925155
    },
    {
      "name": "competitor_spread_ticks:3",
      "parameter": "competitor_spread_ticks",
      "bucket": "3",
      "sample_min": 3,
      "sample_max": 3,
      "attempted": 217,
      "failures": 0,
      "count": 217,
      "mean": 108.35218712042764,
      "se": 11.234423190441198,
      "retail_mean": 132.40297209139308,
      "retail_se": 12.027725220872485,
      "arb_mean": -24.050784971014284,
      "arb_se": 2.501080909761385
    },
    {
      "name": "competitor_spread_ticks:4",
      "parameter": "competitor_spread_ticks",
      "bucket": "4",
      "sample_min": 4,
      "sample_max": 4,
      "attempted": 273,
      "failures": 0,
      "count": 273,
      "mean": 170.61633910404103,
      "se": 13.119139209585898,
      "retail_mean": 196.34209820581538,
      "retail_se": 13.381851480433975,
      "arb_mean": -25.72575910179304,
      "arb_se": 2.7610411990367414
    },
    {
      "name": "retail_arrival_rate:low",
      "parameter": "retail_arrival_rate",
      "bucket": "low",
      "sample_min": 0.15443989293514482,
      "sample_max": 0.21916290059771437,
      "attempted": 334,
      "failures": 0,
      "count": 334,
      "mean": 66.06069019426107,
      "se": 6.2737369801983744,
      "retail_mean": 86.42020577804041,
      "retail_se": 6.760014894332707,
      "arb_mean": -20.359515583832934,
      "arb_se": 2.273779647645315
    },
    {
      "name": "retail_arrival_rate:middle",
      "parameter": "retail_arrival_rate",
      "bucket": "middle",
      "sample_min": 0.2193696884664271,
      "sample_max": 0.28771501984569015,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 106.72358612307897,
      "se": 8.815026425851864,
      "retail_mean": 126.8426360802093,
      "retail_se": 9.66889238901041,
      "arb_mean": -20.119049957123124,
      "arb_se": 1.723666659953332
    },
    {
      "name": "retail_arrival_rate:high",
      "parameter": "retail_arrival_rate",
      "bucket": "high",
      "sample_min": 0.287935927817049,
      "sample_max": 0.35193269318462606,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 122.23680353718468,
      "se": 11.495646077370541,
      "retail_mean": 151.03894018465826,
      "retail_se": 11.721059328742136,
      "arb_mean": -28.80213664745285,
      "arb_se": 3.754917505507159
    },
    {
      "name": "retail_mean_notional:low",
      "parameter": "retail_mean_notional",
      "bucket": "low",
      "sample_min": 2.6452857620506944,
      "sample_max": 3.9262247798901484,
      "attempted": 334,
      "failures": 0,
      "count": 334,
      "mean": 68.11309839704222,
      "se": 6.991826075819572,
      "retail_mean": 86.62213578533203,
      "retail_se": 7.185599545075426,
      "arb_mean": -18.509037388222154,
      "arb_se": 2.2992026797664127
    },
    {
      "name": "retail_mean_notional:middle",
      "parameter": "retail_mean_notional",
      "bucket": "middle",
      "sample_min": 3.9282269322812957,
      "sample_max": 5.094086281721342,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 108.99336651053693,
      "se": 8.913789495439584,
      "retail_mean": 135.55442688347418,
      "retail_se": 9.33360112123878,
      "arb_mean": -26.56106037300901,
      "arb_se": 3.3631708364322597
    },
    {
      "name": "retail_mean_notional:high",
      "parameter": "retail_mean_notional",
      "bucket": "high",
      "sample_min": 5.09968489148911,
      "sample_max": 6.329676759574152,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 117.90845155894925,
      "se": 11.029042213394883,
      "retail_mean": 142.12461297768348,
      "retail_se": 11.775585423772371,
      "arb_mean": -24.216161418756155,
      "arb_se": 2.38462470448063
    },
    {
      "name": "jump_intensity:low",
      "parameter": "jump_intensity",
      "bucket": "low",
      "sample_min": 0.000805597859855233,
      "sample_max": 0.001548516396877783,
      "attempted": 334,
      "failures": 0,
      "count": 334,
      "mean": 107.42737229453653,
      "se": 10.719835892687295,
      "retail_mean": 129.29043309722874,
      "retail_se": 11.450837155634366,
      "arb_mean": -21.863060802670358,
      "arb_se": 2.1035584851617277
    },
    {
      "name": "jump_intensity:middle",
      "parameter": "jump_intensity",
      "bucket": "middle",
      "sample_min": 0.0015493788082731937,
      "sample_max": 0.002335101129155772,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 96.85226553785465,
      "se": 8.808773649577882,
      "retail_mean": 122.04798978767597,
      "retail_se": 9.066331397069737,
      "arb_mean": -25.19572424978018,
      "arb_se": 3.3663725558609974
    },
    {
      "name": "jump_intensity:high",
      "parameter": "jump_intensity",
      "bucket": "high",
      "sample_min": 0.0023356809684818612,
      "sample_max": 0.002999967173767041,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 90.61721775156217,
      "se": 7.803436363319875,
      "retail_mean": 112.83461973662432,
      "retail_se": 8.291882456194783,
      "arb_mean": -22.217401985151053,
      "arb_se": 2.5719183085323443
    },
    {
      "name": "jump_sigma:low",
      "parameter": "jump_sigma",
      "bucket": "low",
      "sample_min": 0.20011205593936066,
      "sample_max": 0.3302169013210605,
      "attempted": 334,
      "failures": 0,
      "count": 334,
      "mean": 117.07727275212515,
      "se": 9.81144848001626,
      "retail_mean": 139.7238287596464,
      "retail_se": 10.531821748659302,
      "arb_mean": -22.646556007462575,
      "arb_se": 1.652254133495155
    },
    {
      "name": "jump_sigma:middle",
      "parameter": "jump_sigma",
      "bucket": "middle",
      "sample_min": 0.3310836477965312,
      "sample_max": 0.46575487961144846,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 94.40147525630871,
      "se": 9.241874873094424,
      "retail_mean": 116.86254303954264,
      "retail_se": 9.515718098762841,
      "arb_mean": -22.461067783286186,
      "arb_se": 3.1009040444450298
    },
    {
      "name": "jump_sigma:high",
      "parameter": "jump_sigma",
      "bucket": "high",
      "sample_min": 0.465884928061146,
      "sample_max": 0.599652885814436,
      "attempted": 333,
      "failures": 0,
      "count": 333,
      "mean": 83.38912889546667,
      "se": 8.396052759913966,
      "retail_mean": 107.55533930383423,
      "retail_se": 8.915464699223444,
      "arb_mean": -24.1662104084,
      "arb_se": 3.1704527174010524
    }
  ],
  "paired": null,
  "failures": [],
  "seconds_per_simulation": 1.046760363716632,
  "methodology": {
    "bucketing": "Marginal buckets, not cross-classified or covariate-adjusted. Continuous parameters sorted by (value, seed); bucket=floor(3*rank/n), zero-based ranks, resulting in 334/333/333. Report actual inclusive sample minima/maxima. Membership computed independently for each parameter across all 1000 seeds.",
    "standard_errors": "Sample standard deviation (ddof=1) / sqrt(count), treating a simulation as the unit. se refers to total edge; retail_se and arb_se use their own simulation-level samples. Contrast SE is sqrt(se_high^2+se_low^2), an unpaired descriptive estimate.",
    "failures": "All attempted seeds are retained for bucket assignment. Means exclude failed simulations, following BatchResult.mean_edge. No failed simulations occurred.",
    "interpretation": "Overlapping marginal partitions are not independent tests. No causal conclusions, multiplicity adjustment, or candidate gain is claimed. Randomly sampled nuisance parameters and simulation paths can affect between-bucket differences.",
    "timing": "Per-simulation elapsed time around run_batch, excluding replay time and parameter extraction. Remaining 980 runs used eight separate OS processes on disjoint seed ranges; each process ran serially with workers=1. Pilot completed before launch.",
    "reproduction": "Put the pinned simulator source on PYTHONPATH. Save code to harness.py and verified incumbent code to B.py. Run python3 harness.py B.py pilot.jsonl 232000 232020 --verify, then python3 harness.py B.py rest.jsonl 232020 233000. Aggregate with python3 harness.py --summarize pilot.jsonl rest.jsonl. Only Python standard-library dependencies. The same aggregate function processes the eight disjoint output files used here."
  },
  "pilot": {
    "n": 20,
    "seconds_per_simulation": 0.9477279120474122,
    "projected_1000_seconds": 947.7279120474122,
    "simulation_budget_seconds": 1800,
    "selected_n": 1000,
    "identical_complete_results": 20,
    "max_abs_total_edge_difference": 0.0,
    "max_abs_retail_edge_difference": 0.0,
    "max_abs_arb_edge_difference": 0.0,
    "instrumentation": "None. Baseline run_batch results were replayed unmodified and compared in full against the results whose regime/edges were extracted. Extraction is post-run only.",
    "rows": [
      {
        "seed": 232000,
        "baseline": [
          241.4890582493,
          259.5536672136,
          -18.0646089636
        ],
        "captured": [
          241.4890582493,
          259.5536672136,
          -18.0646089636
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232001,
        "baseline": [
          32.4222111843,
          35.1635474332,
          -2.7413362495
        ],
        "captured": [
          32.4222111843,
          35.1635474332,
          -2.7413362495
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232002,
        "baseline": [
          380.4633908566,
          403.8806796516,
          -23.4172887945
        ],
        "captured": [
          380.4633908566,
          403.8806796516,
          -23.4172887945
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232003,
        "baseline": [
          35.42501146,
          48.9307914378,
          -13.5057799768
        ],
        "captured": [
          35.42501146,
          48.9307914378,
          -13.5057799768
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232004,
        "baseline": [
          686.8445483705,
          806.0909545482,
          -119.2464061766
        ],
        "captured": [
          686.8445483705,
          806.0909545482,
          -119.2464061766
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232005,
        "baseline": [
          0.3844102166,
          0.619530246,
          -0.2351200299
        ],
        "captured": [
          0.3844102166,
          0.619530246,
          -0.2351200299
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232006,
        "baseline": [
          337.4875533141,
          381.8357730125,
          -44.3482196988
        ],
        "captured": [
          337.4875533141,
          381.8357730125,
          -44.3482196988
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232007,
        "baseline": [
          -243.5939925863,
          46.8300914139,
          -290.4240840001
        ],
        "captured": [
          -243.5939925863,
          46.8300914139,
          -290.4240840001
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232008,
        "baseline": [
          6.196881201,
          7.1787785482,
          -0.981897346
        ],
        "captured": [
          6.196881201,
          7.1787785482,
          -0.981897346
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232009,
        "baseline": [
          58.7281208598,
          69.4294087605,
          -10.7012879018
        ],
        "captured": [
          58.7281208598,
          69.4294087605,
          -10.7012879018
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232010,
        "baseline": [
          125.5596401141,
          151.6020799693,
          -26.0424398561
        ],
        "captured": [
          125.5596401141,
          151.6020799693,
          -26.0424398561
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232011,
        "baseline": [
          6.4600977227,
          17.1404963125,
          -10.6803985901
        ],
        "captured": [
          6.4600977227,
          17.1404963125,
          -10.6803985901
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232012,
        "baseline": [
          4.9857923462,
          5.6403918961,
          -0.6545995494
        ],
        "captured": [
          4.9857923462,
          5.6403918961,
          -0.6545995494
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232013,
        "baseline": [
          5.7359626514,
          13.1998305308,
          -7.4638678797
        ],
        "captured": [
          5.7359626514,
          13.1998305308,
          -7.4638678797
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232014,
        "baseline": [
          50.6235666534,
          61.0552905959,
          -10.4317239414
        ],
        "captured": [
          50.6235666534,
          61.0552905959,
          -10.4317239414
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232015,
        "baseline": [
          20.7172596321,
          26.4645348978,
          -5.7472752663
        ],
        "captured": [
          20.7172596321,
          26.4645348978,
          -5.7472752663
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232016,
        "baseline": [
          -0.1311892882,
          0.1813141792,
          -0.3125034674
        ],
        "captured": [
          -0.1311892882,
          0.1813141792,
          -0.3125034674
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232017,
        "baseline": [
          6.776724451,
          16.746097802,
          -9.9693733515
        ],
        "captured": [
          6.776724451,
          16.746097802,
          -9.9693733515
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232018,
        "baseline": [
          331.7667511031,
          374.0377337208,
          -42.2709826178
        ],
        "captured": [
          331.7667511031,
          374.0377337208,
          -42.2709826178
        ],
        "identical_complete_result": true
      },
      {
        "seed": 232019,
        "baseline": [
          308.9251789936,
          346.9330655912,
          -38.0078865976
        ],
        "captured": [
          308.9251789936,
          346.9330655912,
          -38.0078865976
        ],
        "identical_complete_result": true
      }
    ]
  },
  "contrasts": [
    {
      "parameter": "competitor_spread_ticks",
      "comparison": "4 minus 1",
      "mean_diff": 131.0935780273441,
      "se": 14.87376399009396,
      "retail_diff": 135.74251451723,
      "retail_se": 15.496860323845187,
      "arb_diff": -4.648936489922658,
      "arb_se": 5.108363320786533
    },
    {
      "parameter": "retail_arrival_rate",
      "comparison": "high minus low",
      "mean_diff": 56.176113342923614,
      "se": 13.096169456481284,
      "retail_diff": 64.61873440661785,
      "retail_se": 13.530743998742016,
      "arb_diff": -8.442621063619917,
      "arb_se": 4.3897015114025875
    },
    {
      "parameter": "retail_mean_notional",
      "comparison": "high minus low",
      "mean_diff": 49.79535316190703,
      "se": 13.058537591145374,
      "retail_diff": 55.50247719235145,
      "retail_se": 13.794827033882973,
      "arb_diff": -5.7071240305340005,
      "arb_se": 3.3125168594083236
    },
    {
      "parameter": "jump_intensity",
      "comparison": "high minus low",
      "mean_diff": -16.81015454297436,
      "se": 13.259279793508007,
      "retail_diff": -16.455813360604424,
      "retail_se": 14.137785761292585,
      "arb_diff": -0.35434118248069524,
      "arb_se": 3.3226077238006564
    },
    {
      "parameter": "jump_sigma",
      "comparison": "high minus low",
      "mean_diff": -33.688143856658485,
      "se": 12.913489970696236,
      "retail_diff": -32.16848945581218,
      "retail_se": 13.798723859423829,
      "arb_diff": -1.5196544009374264,
      "arb_se": 3.5751523261152864
    }
  ],
  "provenance": {
    "simulator_commit": "b244530ee2a6247b5f930830eeca9dd950da0e7e",
    "simulator_modified": false,
    "source_urls": [
      "https://api.imdnotdead.fun/artifacts/ecbc9087b4f238fc1b5b0d52f55af073f676c038bfb6756172b1e51480ac1741",
      "https://raw.githubusercontent.com/danrobinson/prediction-market-challenge/b244530ee2a6247b5f930830eeca9dd950da0e7e/orderbook_pm_challenge/runner.py",
      "https://raw.githubusercontent.com/danrobinson/prediction-market-challenge/b244530ee2a6247b5f930830eeca9dd950da0e7e/orderbook_pm_challenge/config.py",
      "https://raw.githubusercontent.com/danrobinson/prediction-market-challenge/b244530ee2a6247b5f930830eeca9dd950da0e7e/orderbook_pm_challenge/results.py",
      "https://raw.githubusercontent.com/danrobinson/prediction-market-challenge/b244530ee2a6247b5f930830eeca9dd950da0e7e/docs/orderbook_prediction_market_challenge.md",
      "https://raw.githubusercontent.com/danrobinson/prediction-market-challenge/b244530ee2a6247b5f930830eeca9dd950da0e7e/orderbook_pm_challenge/engine.py"
    ],
    "measurement_source": "All numerical estimates, counts and verification outcomes are produced locally by the supplied harness on the specified seeds, not claimed as published source facts.",
    "raw_records_sha256": "34126979a9458439f1f41007c971d34605b6e6225f698f80a5d59e016a76330b"
  },
  "execution": {
    "remaining_980_wall_seconds": 130.7955290209502
  }
}