# Remove-current merge 02: one-deletion screening

The evidence below is five **separate policies** on screening seeds 410000–410499, with signed differences `d = candidate total_edge − reported control total_edge`. `se` is each source's paired standard error; `retail_diff` and `arb_diff` are signed candidate-minus-control means. The JSON table preserves each source's reported `n`, `sum_d`, `sum_d2`, `mean_diff`, `se`, and retail/arb differences as supplied, including their decimal precision. Source `complete` records the item's own status; the flags separately record comparability and protocol issues. Each `code_url` is the supplied item-result artifact containing `JSON.code`, not a separate raw-code artifact. The five supplied result SHA-256 values and their embedded `code_sha256` values were verified against the bytes read for this merge. The named [incumbent B artifact](https://api.imd.fun/artifacts/db22aff7167a4798251ed3d20432509f30e511e4958799aba5b17b41bb39b4b6) yielded the stated B code hash `bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae`; the four non-hysteresis candidates were compared bytewise with that code. The [combine-cash artifact](https://api.imd.fun/artifacts/ecbc9087b4f238fc1b5b0d52f55af073f676c038bfb6756172b1e51480ac1741) yielded `baef955ec10728839f7c3357016ec9eb0550e7ec689bbd5171b1cdd5c9e37273`, matching hysteresis's reported paired reference. These code checks establish identity and edit scope; they do not independently reproduce the simulations.

## JSON evidence table

```json
{
  "rows": [
    {
      "key": "remove-flow-quote",
      "idea": "One deletion, without tuning: in the existing self.s == 1 qsize branch replace learned rr/mm with rr=rho and mm=mu, using the fixed retail prior (rho=0.125, mean notional=5.0, size sigma=1.2). All other source bytes, branches, parameters, flow-aware stock demand and estimator are preserved.",
      "target_spreads": [
        1
      ],
      "input_result_locator": "artifacts/goal/batches/r3-remove-current/items/flow-quote/result.json",
      "read_locator": ".imd/reads/artifacts/item-remove-flow-quote",
      "artifact_sha256": "d8509b85a87c868fccfdb32e87210b401d02c296ede52a6b8a67633ace7d22a1",
      "code_url": "https://api.imd.fun/artifacts/d8509b85a87c868fccfdb32e87210b401d02c296ede52a6b8a67633ace7d22a1",
      "code_sha256": "9ff5bc5e908ed84b50bfde5cad8be70eb9e17c89a2d76001af0c88dbbb023572",
      "incumbent_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "paired_comparator_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "complete": true,
      "failures": [],
      "tuning": null,
      "paired": {
        "seed_start": 410000,
        "seed_end_inclusive": 410499,
        "statistics": {
          "overall": {
            "n": 500,
            "sum_d": -82.34618453250002,
            "sum_d2": 3777.4410782025857,
            "mean_diff": -0.16469236906500004,
            "se": 0.12282381095303921,
            "C_mean_edge": 104.2490608190818,
            "B_mean_edge": 104.4137531881468,
            "retail_diff": 0.0186156085948004,
            "arb_diff": -0.18330797765199994
          },
          "s1": {
            "n": 115,
            "sum_d": -82.34618453250002,
            "sum_d2": 3777.4410782025857,
            "mean_diff": -0.7160537785434784,
            "se": 0.5325755090993263,
            "C_mean_edge": 27.779100681187828,
            "B_mean_edge": 28.495154459731307,
            "retail_diff": 0.08093742867304521,
            "arb_diff": -0.7969912071826085
          },
          "s2": {
            "n": 116,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 97.76626247947758,
            "B_mean_edge": 97.76626247947758,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s3": {
            "n": 138,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 128.4054514027384,
            "B_mean_edge": 128.4054514027384,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s4": {
            "n": 131,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 151.67248160310686,
            "B_mean_edge": 151.67248160310686,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          }
        }
      },
      "seconds_per_simulation": {
        "all": 0.5037353071724064,
        "B": 0.5037964236158877,
        "C": 0.5036741907289252,
        "n_simulations": 1000,
        "method": "perf_counter around SimulationEngine(config, fresh_observing_strategy_factory, seed=seed).run(); includes construction/run and first-state observation wrapper; first 10 pairs retained exactly once in paired.first_10_timed_pairs."
      },
      "flags": {
        "input_artifact_sha256_verified": true,
        "candidate_code_sha256_verified": true,
        "declared_B_hash_matches_verified_incumbent_artifact": true,
        "paired_comparator_mismatch": false,
        "missing_pair_count": 0,
        "protocol_violations": [],
        "spread_label_verification": "reported 0 mismatches against sampled config and result regime",
        "reported_failed_flags": 0,
        "cross_item_B_baseline_discrepancy": false,
        "finalist_seeds_measured": false,
        "eligible_for_promotion": false,
        "source_spread_mismatch_count": 0
      }
    },
    {
      "key": "remove-flow-side-pool",
      "idea": "Pure deletion from incumbent B (spread 4 only, inside the existing `elif self.s == 4` branch): the end-of-step loop that registers flow observations for the pooled f4 estimator (f4_exposure/f4_events/f4_weights -> f4_rate, f4_mean, f4_flow) now iterates only the ask side (side 0); bid-side (side 1) observations are no longer recorded. Ask-side likelihood, priors (80-exposure rate prior, 9-point mean grid), the 25-step refresh and all downstream uses (qask, ammo/flow_scale, ydf1 stock_rate) are unchanged. One-line diff at line 927: tuple `((0, ask_t, ...), (1, bid_t, cb, max(phat, 0.05), phat - bid_t / 100.0))` -> `((0, ask_t, ...),)`. No tuning.",
      "target_spreads": [
        4
      ],
      "input_result_locator": "artifacts/goal/batches/r3-remove-current/items/flow-side-pool/result.json",
      "read_locator": ".imd/reads/artifacts/item-remove-flow-side-pool",
      "artifact_sha256": "0846b1a646fcc73ecb77175f9a5349f4e0c0f4f99a10972f73f44c2bb52a04ac",
      "code_url": "https://api.imd.fun/artifacts/0846b1a646fcc73ecb77175f9a5349f4e0c0f4f99a10972f73f44c2bb52a04ac",
      "code_sha256": "2a02e5179256887f67ccdd09281a2982baa373e4a4648d9d20768ee70873a6e2",
      "incumbent_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "paired_comparator_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "complete": true,
      "failures": {
        "missing_seeds": [],
        "failed_simulations": [],
        "exceptions": [],
        "failed_flags": 0,
        "spread_label_mismatches": [],
        "label_method": "s read from strategy instance (self.s = max(1,(ca-cb+1)//2) from first observed rival quotes) after each run; checked equal to sampled regime.competitor_spread_ticks for all 1000 runs"
      },
      "tuning": {
        "performed": false,
        "reason": "Brief specifies no tuning (null); no tuning seeds were run.",
        "n": 0
      },
      "paired": {
        "seed_start": 410000,
        "seed_end_inclusive": 410499,
        "statistics": {
          "overall": {
            "n": 500,
            "sum_d": -3.1470619674,
            "sum_d2": 31878.1622391763,
            "mean_diff": -0.0062941239,
            "se": 0.3574466477,
            "C_mean_edge": 104.4074590642,
            "B_mean_edge": 104.4137531881,
            "retail_diff": 0.0857803874,
            "arb_diff": -0.0920745114,
            "n_nonzero_d": 131
          },
          "s1": {
            "n": 115,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 28.4951544597,
            "B_mean_edge": 28.4951544597,
            "retail_diff": 0.0,
            "arb_diff": 0.0,
            "n_nonzero_d": 0
          },
          "s2": {
            "n": 116,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 97.7662624795,
            "B_mean_edge": 97.7662624795,
            "retail_diff": 0.0,
            "arb_diff": 0.0,
            "n_nonzero_d": 0
          },
          "s3": {
            "n": 138,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 128.4054514027,
            "B_mean_edge": 128.4054514027,
            "retail_diff": 0.0,
            "arb_diff": 0.0,
            "n_nonzero_d": 0
          },
          "s4": {
            "n": 131,
            "sum_d": -3.1470619674,
            "sum_d2": 31878.1622391763,
            "mean_diff": -0.0240233738,
            "se": 1.3681660178,
            "C_mean_edge": 151.6484582293,
            "B_mean_edge": 151.6724816031,
            "retail_diff": 0.3274060589,
            "arb_diff": -0.3514294327,
            "n_nonzero_d": 131
          }
        }
      },
      "seconds_per_simulation": 1.3912,
      "flags": {
        "input_artifact_sha256_verified": true,
        "candidate_code_sha256_verified": true,
        "declared_B_hash_matches_verified_incumbent_artifact": true,
        "paired_comparator_mismatch": false,
        "missing_pair_count": 0,
        "protocol_violations": [
          "Three resumed OS process chunks; item required one OS process."
        ],
        "spread_label_verification": "reported 0 mismatches against sampled config and result regime",
        "reported_failed_flags": 0,
        "cross_item_B_baseline_discrepancy": false,
        "finalist_seeds_measured": false,
        "eligible_for_promotion": false,
        "source_reported_nonzero_d": {
          "overall": 131,
          "s1": 0,
          "s2": 0,
          "s3": 0,
          "s4": 131
        },
        "source_spread_mismatch_count": 0
      }
    },
    {
      "key": "remove-hysteresis",
      "idea": "One deletion from combine-cash: remove only the keep_lo replenishment deadband, so positive target-size deficits are added immediately; retention, caps, budgets, and all parameters are unchanged.",
      "target_spreads": [
        3
      ],
      "input_result_locator": "artifacts/goal/batches/r3-remove-current/items/hysteresis/result.json",
      "read_locator": ".imd/reads/artifacts/item-remove-hysteresis",
      "artifact_sha256": "ecc23ad790caebd237f82737003d42f89b40fb82e730fe36eaa0a315cd6a6351",
      "code_url": "https://api.imd.fun/artifacts/ecc23ad790caebd237f82737003d42f89b40fb82e730fe36eaa0a315cd6a6351",
      "code_sha256": "7102492c3bd187b2c6a7dd6324a25b8ab2c2f115c784f8439ceac71f05446587",
      "incumbent_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "paired_comparator_code_sha256": "baef955ec10728839f7c3357016ec9eb0550e7ec689bbd5171b1cdd5c9e37273",
      "complete": true,
      "failures": {
        "failed_seeds": [],
        "exceptions": [],
        "note": "The named incumbent artifact URL returned HTTP 404, so the numeric pair is against the available combine-cash artifact (reference_code_sha256), not verified hash bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae."
      },
      "tuning": null,
      "paired": {
        "seed_start": 410000,
        "seed_end_inclusive": 410499,
        "statistics": {
          "overall": {
            "n": 500,
            "sum_d": 445.4879019535997,
            "sum_d2": 39577.6536320857,
            "mean_diff": 0.8909758039071994,
            "se": 0.39627895829531734,
            "C_mean_edge": 105.74637229335036,
            "B_mean_edge": 104.85539648944311,
            "retail_diff": 1.7296931868136018,
            "arb_diff": -0.8387173829191995,
            "C_mean_retail": 131.7287143521727,
            "B_mean_retail": 129.9990211653593,
            "C_mean_arb": -25.982342058843614,
            "B_mean_arb": -25.143624675924393
          },
          "s1": {
            "n": 115,
            "sum_d": 162.62221948480015,
            "sum_d2": 13740.525514021601,
            "mean_diff": 1.4141062563895666,
            "se": 1.0151619612045901,
            "C_mean_edge": 28.643896532454793,
            "B_mean_edge": 27.229790276065234,
            "retail_diff": 0.8704315217365217,
            "arb_diff": 0.543674734649565,
            "C_mean_retail": 45.99417596872435,
            "B_mean_retail": 45.12374444698782,
            "C_mean_arb": -17.350279436233915,
            "B_mean_arb": -17.89395417088347
          },
          "s2": {
            "n": 116,
            "sum_d": -11.20183419179998,
            "sum_d2": 6700.559122563973,
            "mean_diff": -0.09656753613620672,
            "se": 0.7086675520496376,
            "C_mean_edge": 96.3870285876931,
            "B_mean_edge": 96.48359612382932,
            "retail_diff": 1.4367473095370709,
            "arb_diff": -1.5333148457137928,
            "C_mean_retail": 129.6134188360777,
            "B_mean_retail": 128.17667152654053,
            "C_mean_arb": -33.2263902484431,
            "B_mean_arb": -31.69307540272933
          },
          "s3": {
            "n": 138,
            "sum_d": 72.24782842459987,
            "sum_d2": 6337.77415156103,
            "mean_diff": 0.523534988584057,
            "se": 0.5772563521459559,
            "C_mean_edge": 129.623677567584,
            "B_mean_edge": 129.10014257899994,
            "retail_diff": 1.538166462145652,
            "arb_diff": -1.0146314735797108,
            "C_mean_retail": 154.89600501775732,
            "B_mean_retail": 153.35783855561152,
            "C_mean_arb": -25.27232745023623,
            "B_mean_arb": -24.257695976656525
          },
          "s4": {
            "n": 131,
            "sum_d": 221.81968823599973,
            "sum_d2": 12798.794843939118,
            "mean_diff": 1.693280062870227,
            "se": 0.8541010867402841,
            "C_mean_edge": 156.56622309117478,
            "B_mean_edge": 154.8729430283046,
            "retail_diff": 2.9451702956083983,
            "arb_diff": -1.251890232729008,
            "C_mean_retail": 184.45970734540157,
            "B_mean_retail": 181.514537049793,
            "C_mean_arb": -27.89348425422061,
            "B_mean_arb": -26.641594021491606
          }
        }
      },
      "seconds_per_simulation": 1.117657,
      "flags": {
        "input_artifact_sha256_verified": true,
        "candidate_code_sha256_verified": true,
        "declared_B_hash_matches_verified_incumbent_artifact": true,
        "paired_comparator_mismatch": true,
        "missing_pair_count": 0,
        "protocol_violations": [
          "Paired control is combine-cash, not incumbent B.",
          "Two separate CLI runs; item required one OS process.",
          "Candidate is not a one-deletion edit of incumbent B."
        ],
        "spread_label_verification": "regime spread only; no observed-quote cross-check reported",
        "reported_failed_flags": null,
        "cross_item_B_baseline_discrepancy": false,
        "finalist_seeds_measured": false,
        "eligible_for_promotion": false,
        "source_spread_mismatch_count": null
      }
    },
    {
      "key": "remove-late-uncovered",
      "idea": "One deletion without tuning: suppress only terminal H<Hu and phat<0.05 uncovered-ask budget override for self.s==4. Ordinary fu/xu collateral policy and final-step pricing retained.",
      "target_spreads": [
        4
      ],
      "input_result_locator": "artifacts/goal/batches/r3-remove-current/items/late-uncovered/result.json",
      "read_locator": ".imd/reads/artifacts/item-remove-late-uncovered",
      "artifact_sha256": "a878833b13826f89f3791f14242315056c08857e3e81684c92bbdcc97d40e121",
      "code_url": "https://api.imd.fun/artifacts/a878833b13826f89f3791f14242315056c08857e3e81684c92bbdcc97d40e121",
      "code_sha256": "125a781db0303d1d5b478a53aa9fcdeb22cb96f9ef0ea4bf40a12bf2d0491a04",
      "incumbent_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "paired_comparator_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "complete": false,
      "failures": [],
      "tuning": null,
      "paired": {
        "seed_start": 410000,
        "seed_end_inclusive": 410499,
        "statistics": {
          "overall": {
            "n": 222,
            "sum_d": -35.63870507119999,
            "sum_d2": 405.18009639406193,
            "mean_diff": -0.1605347075279279,
            "se": 0.09023258476314784,
            "C_mean_edge": 106.34622340200451,
            "B_mean_edge": 106.50675810953243,
            "retail_diff": -0.15963684903288286,
            "arb_diff": -0.0008978584968468446
          },
          "s1": {
            "n": 48,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 23.591521670116666,
            "B_mean_edge": 23.591521670116666,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s2": {
            "n": 56,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 94.07241753911964,
            "B_mean_edge": 94.07241753911964,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s3": {
            "n": 58,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 139.00274293063103,
            "B_mean_edge": 139.00274293063103,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s4": {
            "n": 60,
            "sum_d": -35.63870507119999,
            "sum_d2": 405.18009639406193,
            "mean_diff": -0.5939784178533333,
            "se": 0.3293597087131268,
            "C_mean_edge": 152.437568048535,
            "B_mean_edge": 153.03154646638833,
            "retail_diff": -0.5906563414216666,
            "arb_diff": -0.003322076438333325
          }
        },
        "completed_seed_end_inclusive": 410221,
        "missing_seed_range_inclusive": [
          410222,
          410499
        ]
      },
      "seconds_per_simulation": {
        "mean": 6.086456236024419,
        "B_mean": 6.122076772861308,
        "C_mean": 6.0508356991875285,
        "timed_simulations": 444,
        "elapsed_seconds": 2702.6871017380618
      },
      "flags": {
        "input_artifact_sha256_verified": true,
        "candidate_code_sha256_verified": true,
        "declared_B_hash_matches_verified_incumbent_artifact": true,
        "paired_comparator_mismatch": false,
        "missing_pair_count": 278,
        "protocol_violations": [],
        "spread_label_verification": "reported verified for 222 pairs; mismatch count not separately supplied",
        "reported_failed_flags": 0,
        "cross_item_B_baseline_discrepancy": false,
        "finalist_seeds_measured": false,
        "eligible_for_promotion": false,
        "source_spread_mismatch_count": null
      }
    },
    {
      "key": "remove-probe-renewal",
      "idea": "One deletion from B, guarded by self.s == 3: retain one live probe whose side and price remain intended; track its price in pb/pa and skip duplicate probe placement. Cancel stale/duplicate probes; cancel retained probes if the original free>1 placement gate closes. Other branches and parameters preserved.",
      "target_spreads": [
        3
      ],
      "input_result_locator": "artifacts/goal/batches/r3-remove-current/items/probe-renewal/result.json",
      "read_locator": ".imd/reads/artifacts/item-remove-probe-renewal",
      "artifact_sha256": "b52dfb9c23cdf3787f6f9c0a100eeab9f0c08e95b96a48e28c4d6acf2f1148fb",
      "code_url": "https://api.imd.fun/artifacts/b52dfb9c23cdf3787f6f9c0a100eeab9f0c08e95b96a48e28c4d6acf2f1148fb",
      "code_sha256": "bb316faba7036273cd1c6a79c746c8d8a68aab0257bc81948793930829303fdc",
      "incumbent_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "paired_comparator_code_sha256": "bf0d179f3e47d28819f6fa7e0d25548c2ced43374c216d36460b00a26a26ffae",
      "complete": true,
      "failures": [],
      "tuning": null,
      "paired": {
        "seed_start": 410000,
        "seed_end_inclusive": 410499,
        "statistics": {
          "overall": {
            "n": 500,
            "sum_d": 131.3004009240001,
            "sum_d2": 16185.199087094086,
            "mean_diff": 0.26260080184800016,
            "se": 0.25442551497728527,
            "C_mean_edge": 104.65477663116309,
            "B_mean_edge": 104.39217582931512,
            "retail_diff": 0.20281271113859975,
            "arb_diff": 0.05978809070920001
          },
          "s1": {
            "n": 115,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 28.495164940043484,
            "B_mean_edge": 28.495164940043484,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s2": {
            "n": 116,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 97.76628686140519,
            "B_mean_edge": 97.76628686140519,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          },
          "s3": {
            "n": 138,
            "sum_d": 131.3004009240001,
            "sum_d2": 16185.199087094086,
            "mean_diff": 0.9514521806086963,
            "se": 0.921672512194837,
            "C_mean_edge": 129.3571080834449,
            "B_mean_edge": 128.40565590283617,
            "retail_diff": 0.7348286635456512,
            "arb_diff": 0.2166235170623189
          },
          "s4": {
            "n": 131,
            "sum_d": 0.0,
            "sum_d2": 0.0,
            "mean_diff": 0.0,
            "se": 0.0,
            "C_mean_edge": 151.58987905372672,
            "B_mean_edge": 151.58987905372672,
            "retail_diff": 0.0,
            "arb_diff": 0.0
          }
        }
      },
      "seconds_per_simulation": 1.1841114881578834,
      "flags": {
        "input_artifact_sha256_verified": true,
        "candidate_code_sha256_verified": true,
        "declared_B_hash_matches_verified_incumbent_artifact": true,
        "paired_comparator_mismatch": false,
        "missing_pair_count": 0,
        "protocol_violations": [],
        "spread_label_verification": "reported 0 mismatches against sampled config and result regime",
        "reported_failed_flags": 0,
        "cross_item_B_baseline_discrepancy": true,
        "finalist_seeds_measured": false,
        "eligible_for_promotion": false,
        "largest_absolute_pair": {
          "seed": 410361,
          "s": 3,
          "d": 127.20638946370002,
          "C": {
            "total_edge": 533.0016783459,
            "retail_edge": 666.1402194642,
            "arb_edge": -133.1385411186,
            "failed": false,
            "error": null
          },
          "B": {
            "total_edge": 405.7952888822,
            "retail_edge": 564.6600126213,
            "arb_edge": -158.8647237397,
            "failed": false,
            "error": null
          }
        },
        "source_spread_mismatch_count": 0
      }
    }
  ]
}
```

## Screening by spread

| Candidate (C − reported control) | s=1 gain ± SE (n) | s=2 gain ± SE (n) | s=3 gain ± SE (n) | s=4 gain ± SE (n) | Overall gain ± SE (n) |
|---|---:|---:|---:|---:|---:|
| [remove-flow-quote](https://api.imd.fun/artifacts/d8509b85a87c868fccfdb32e87210b401d02c296ede52a6b8a67633ace7d22a1) | -0.7160537785434784 ± 0.5325755090993263 (115) | 0.0 ± 0.0 (116) | 0.0 ± 0.0 (138) | 0.0 ± 0.0 (131) | -0.16469236906500004 ± 0.12282381095303921 (500) |
| [remove-flow-side-pool](https://api.imd.fun/artifacts/0846b1a646fcc73ecb77175f9a5349f4e0c0f4f99a10972f73f44c2bb52a04ac) | 0.0 ± 0.0 (115) | 0.0 ± 0.0 (116) | 0.0 ± 0.0 (138) | -0.0240233738 ± 1.3681660178 (131) | -0.0062941239 ± 0.3574466477 (500) |
| [remove-hysteresis](https://api.imd.fun/artifacts/ecc23ad790caebd237f82737003d42f89b40fb82e730fe36eaa0a315cd6a6351) | 1.4141062563895666 ± 1.0151619612045901 (115) | -0.09656753613620672 ± 0.7086675520496376 (116) | 0.523534988584057 ± 0.5772563521459559 (138) | 1.693280062870227 ± 0.8541010867402841 (131) | 0.8909758039071994 ± 0.39627895829531734 (500) |
| [remove-late-uncovered](https://api.imd.fun/artifacts/a878833b13826f89f3791f14242315056c08857e3e81684c92bbdcc97d40e121) | 0.0 ± 0.0 (48) | 0.0 ± 0.0 (56) | 0.0 ± 0.0 (58) | -0.5939784178533333 ± 0.3293597087131268 (60) | -0.1605347075279279 ± 0.09023258476314784 (222) |
| [remove-probe-renewal](https://api.imd.fun/artifacts/b52dfb9c23cdf3787f6f9c0a100eeab9f0c08e95b96a48e28c4d6acf2f1148fb) | 0.0 ± 0.0 (115) | 0.0 ± 0.0 (116) | 0.9514521806086963 ± 0.921672512194837 (138) | 0.0 ± 0.0 (131) | 0.26260080184800016 ± 0.25442551497728527 (500) |

These are separate paired comparisons. In particular, hysteresis uses combine-cash as its control, and late-uncovered covers only the first 222 seeds. No means or sums from different candidates or controls have been pooled. The zero differences outside the stated target spread for flow-quote, side-pool, late-uncovered and probe-renewal are present in their source statistics; hysteresis shows nonzero differences in all four spreads. See each linked row above.

## Findings and disposition

- **Flow quote, s=1:** replacing the learned flow rate and size estimates with fixed retail priors changes only the s=1 quote-size calculation in B. Its 500-pair screen is negative overall and at s=1; the s=1 retail difference is positive while the arbitrage difference is more negative. The candidate code differs from B only at the two reported assignments. Source: [flow-quote result](https://api.imd.fun/artifacts/d8509b85a87c868fccfdb32e87210b401d02c296ede52a6b8a67633ace7d22a1).
- **Flow side pool, s=4:** omitting bid-side observations from the pooled flow estimator changes one loop tuple in B. The s=4 effect is near zero relative to its large SE, so this screen does not resolve whether the bid-side data help. The source reports 500 paired seeds, but its reproduction describes three resumed single-process chunks, violating the item's one-OS-process requirement. Source: [side-pool result](https://api.imd.fun/artifacts/0846b1a646fcc73ecb77175f9a5349f4e0c0f4f99a10972f73f44c2bb52a04ac).
- **Hysteresis, nominal s=3:** removing the `keep_lo` replenishment deadband is a small edit to combine-cash, not to incumbent B. Its reported +0.8909758039071994 overall gain (SE 0.39627895829531734) is therefore **not a gain over B**. The source ran the two policies through separate CLI invocations, also violating the one-OS-process requirement. The source's `complete=true` denotes its 500 completed pairs only. Sources: [hysteresis result](https://api.imd.fun/artifacts/ecc23ad790caebd237f82737003d42f89b40fb82e730fe36eaa0a315cd6a6351), [combine-cash control](https://api.imd.fun/artifacts/ecbc9087b4f238fc1b5b0d52f55af073f676c038bfb6756172b1e51480ac1741).
- **Late uncovered ask, s=4:** suppressing the terminal low-probability uncovered-ask budget override is a one-condition edit of B. The source stopped at seed 410221 after 222 of 500 pairs; 278 seeds (410222–410499) are missing. Its s=4 loss is mainly retail rather than arbitrage. The source correctly marks `complete=false`; this is an incomplete screen. Source: [late-uncovered result](https://api.imd.fun/artifacts/a878833b13826f89f3791f14242315056c08857e3e81684c92bbdcc97d40e121).
- **Probe renewal, s=3:** retaining correctly priced live probes removes repeat placement while cancelling stale or duplicate probes. Its positive s=3 screen has SE nearly as large as its mean. The source identifies seed 410361 as the largest pair (`d=127.20638946370002`) against overall `sum_d=131.3004009240001`; the result is dominated by that pair. Its reported B mean on the same 500 seeds differs from the flow-quote and side-pool reports even though they name the same B code hash. The cause cannot be determined from these result objects, so the baseline discrepancy remains flagged. Sources: [probe-renewal result](https://api.imd.fun/artifacts/b52dfb9c23cdf3787f6f9c0a100eeab9f0c08e95b96a48e28c4d6acf2f1148fb), [flow-quote result](https://api.imd.fun/artifacts/d8509b85a87c868fccfdb32e87210b401d02c296ede52a6b8a67633ace7d22a1), [side-pool result](https://api.imd.fun/artifacts/0846b1a646fcc73ecb77175f9a5349f4e0c0f4f99a10972f73f44c2bb52a04ac).

**No promotion.** All reported numbers are screening evidence from at most 500 pairs, with only 60–138 observations per spread in the affected branches. No supplied result reports finalist-seed measurements on 900000–903999. Final ranking under the stated rule requires thousands of paired seeds against the same verified incumbent, and a gain above twice its finalist standard error; the current evidence does not meet that gate. The next measurement should resolve the flagged protocol and baseline discrepancies, finish missing pairs where relevant, and then test qualified frozen candidates on the reserved finalist seeds. Sources: the five linked item results and the goal's promotion rule in the task brief.