# Merge 14 of r7-t03-joint: items j131–j140 (continuing T3 with CMA-ES) All 10 inputs were read from `.imd/reads/artifacts/item-j131` … `item-j140`. For each one, the SHA-256 of the file matched its id: `sha256sum` gave the `result_sha256` values shown in the table. In every item, `code_zb64` was decoded (base64 then zlib) and its SHA-256 matched the item's `code_sha256`. These checks were run in this merge. The other numbers here are copied from the item results, unless they are marked *recomputed here*. The `sum` and `sumsq` values were used to recompute each block's mean and se, and every one agreed with the item's own figures to within floating-point rounding. ## Summary - **Only 2 of the 10 items ran a search: j136 and j140.** The other 8 measured a replica speed below the 500 markets/s threshold, so they took the brief's real-only branch (T3 against HA on the real engine). One of those 8, j135, has no paired data at all (see below). - **Best valid search row: j136** (es_seed 936, sigma0 0.5, 512 markets/gen, 287 generations, 56.27 min). Holdout over T was **+2.5621 se 0.0136**, and over T3 (the base) **+0.0290 se 0.0043**, on 8192 replica seeds 40000000..40008191. On the real engine, candidate minus T3 was +0.0545 se 0.0123 (seeds 9240000..9241023) and +0.0240 se 0.0142 (9242000..9243023). *Pooled here:* +0.0392 se 0.0094, n=2048. afterSwap peak 237,706 with 0 OOG. Its gate C max was 6.29e-05 bps. - **j140** is also valid but scored slightly lower. Settings: es_seed 940, sigma0 0.3, 1024 markets/gen, 168 generations, 45.02 min. Holdout over T was +2.5559 se 0.0135, and over T3 +0.0228 se 0.0037. Real engine, candidate minus T3: +0.0210 se 0.0104 and +0.0295 se 0.0098, pooled +0.0253 se 0.0072. afterSwap peak 237,771 with 0 OOG. - **Conclusion: both continuations of T3 gain only about +0.02 to +0.05 over T3 on the real engine.** Both replica predictions fall within that range, so the replica stays calibrated, but these curves have flattened: j136's monitor went from +2.5846 to a best of +2.6138, and j140's from +2.5846 to +2.6115. T3's own search curve had still been rising. j140 states the same thing in its notes: *"gains from further CMA on these coordinates at 1024 markets/generation appear to have flattened, unlike T3's curve."* - **T3 minus HA on the real engine** (7 items, 8 blocks, all fresh and non-overlapping seed ranges): every block lies between +1.403 and +1.454. *Pooled here:* **+1.4277 se 0.0106, n=7616**. That puts j136's candidate at about +1.47 over HA, an inference from adding the two pooled differences, not a direct measurement. The goal says the top 10 needs about +3.1 over HA, so roughly +1.6 is still missing. No leaderboard placement is claimed. - **Gas:** no arm in any item hit OOG. The highest afterSwap peak was 237,771 (j140's candidate), which is under the 245,000 limit. ## Per-item rows Holdout values are replica numbers. "Real cand−T3" applies only to search rows. In real-only rows the candidate is T3 itself, so there is no candidate−T3 difference, and those items' T3−HA results are under `real_only`. | key | mode | markets/s | gens | search min | holdout over_T (se) | holdout over_base (se) | real cand−T3 per block (se) | afterSwap peak | OOG | valid | last curve point (min, gain, sigma) | |---|---|---|---|---|---|---|---|---|---|---|---| | j131 | real-only | 172.7 | — | — | — | — | — | 237600 | 0 | true | — | | j132 | real-only | 227.9 | — | — | — | — | — | 237645 | 0 | true | — | | j133 | real-only | 173.3 | — | — | — | — | — | 237587 | 0 | true | — | | j134 | real-only | 252.9 | — | — | +2.5331 (0.0135) | +0.0000 (0.0000) | — | 237642 | 0 | true | — | | j135 | real-only | 472.8 | — | — | — | — | — | 232525 | 0 | false | — | | j136 | search | 1103.2 | 287 | 56.27 | +2.5621 (0.0136) | +0.0290 (0.0043) | 9240000..9241023: +0.0545 (0.0123); 9242000..9243023: +0.0240 (0.0142) | 237706 | 0 | true | 55.4551, 2.6058, 0.1913 | | j137 | real-only | 426.5 | — | — | — | — | — | 237512 | 0 | true | — | | j138 | real-only | 214.0 | — | — | — | — | — | 237600 | 0 | true | — | | j139 | real-only | 202.2 | — | — | — | — | — | 237461 | 0 | true | — | | j140 | search | 957.7 | 168 | 45.02 | +2.5559 (0.0135) | +0.0228 (0.0037) | 9256000..9257023: +0.0210 (0.0104); 9258000..9259023: +0.0295 (0.0098) | 237771 | 0 | true | 45.0500, 2.6115, 0.1375 | Every real-only row has `code_sha256` `cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9`, which is T3 unchanged. In all 8 real-only items the params are equal in value to T3's (checked here). j136 has `f65ab9b7d79a3bc626b685da15847db13f804b57338ec3ca6742ca5315e96fb2` and j140 has `807b2175998dd629761f78ff9f817cb20d0a7af60ffd1d2af5f9e8bea94ed93e`. Both keep T3's scaling. The rounded constants that changed are STEP_COUNT_CAP 64→66 and ANCHOR_COUNT_CAP 715→710 in j136, and ANCHOR_COUNT_CAP 715→704 in j140. ## Real-only: T3 minus HA (each strategy alone against the 30 bps Vanilla) | key | seeds | n | mean | se | sum | sumsq | T3 afterSwap peak / OOG | HA afterSwap peak / OOG | |---|---|---|---|---|---|---|---|---| | j131 | 9220000..9220447 | 448 | +1.4415 | 0.0483 | 645.7905 | 1398.9956 | 237600 / 0 | 231567 / 0 | | j132 | 9224000..9225023 | 1024 | +1.4527 | 0.0299 | 1487.5701 | 3099.9376 | 237645 / 0 | 233576 / 0 | | j133 | 9228000..9229023 | 1024 | +1.4279 | 0.0302 | 1462.1724 | 3040.4897 | 237587 / 0 | 233659 / 0 | | j134 | 9232000..9233023 | 1024 | +1.4206 | 0.0268 | 1454.6846 | 2819.3699 | 237642 / 0 | 233693 / 0 | | j134 | 9234000..9235023 | 1024 | +1.4094 | 0.0281 | 1443.2419 | 2863.1212 | 237642 / 0 | 233693 / 0 | | j135 | requested 9236000..9237023 | 0 | — | — | — | — | 232525 / 0 (8-market probe only) | not run | | j137 | 9244000..9245023 | 1024 | +1.4031 | 0.0269 | 1436.8145 | 2772.3602 | 237512 / 0 | 233591 / 0 | | j138 | 9248000..9249023 | 1024 | +1.4541 | 0.0276 | 1488.9659 | 2965.2825 | 237600 / 0 | 231592 / 0 | | j139 | 9252000..9253023 | 1024 | +1.4202 | 0.0305 | 1454.2669 | 3037.9705 | 237461 / 0 | 231504 / 0 | | **pooled (here)** | 8 blocks | 7616 | +1.4277 | 0.0106 | 10873.5068 | 21997.5273 | max 237645 / 0 | max 233693 / 0 | ## What failed or is incomplete - **j135 (valid=false, blocked at the real-engine step).** Replica speed was 472.8 markets/s on 16 cores, under 500, so it took the real-only branch. Two segments ran in parallel, T3 on 9236008..9236255 and HA on 9236000..9236255, and both hit the 470 s command cap (exit 124). Only an 8-market T3 probe finished, so there are no paired seeds and no HA gas. The item says the cause was not diagnosed. No numbers from it are used here. - **j131 is partial.** It finished 448 of the 1024 requested seeds (9220000..9220447) because the machine was under load average ~45, and it ran no second block. Its +1.4415 se 0.0483 is consistent with the other blocks. Pooling without j131 gives +1.4269 se 0.0108, n=7168 (*recomputed here*). - **8 of 10 items did no search.** Their measured replica speeds were 172.7 (j131), 227.9 (j132), 173.3 (j133), 252.9 (j134), 472.8 (j135), 426.5 (j137), 214.0 (j138) and 202.2 (j139) markets/s. The two search items measured 1103.2 (j136) and 957.7 (j140). Since all items share the same port and replica, the difference comes from the machines. j138 states this directly: *"the difference is the hardware, not the code"*. As a result, the settings these items were assigned were never tested. The recorded assignments are: j132 sigma0 0.5, es 932; j133 setting D, sigma0 0.3, 1024/gen, es 933; j134 sigma0 0.5, es 934; j135 setting B, sigma0 1.0, es 935; j137 sigma0 1.0, es 937; j139 setting A, sigma0 0.5, es 939. j131 and j138 record no assigned setting. - **No item could provide retail/arb edge deltas on the real engine.** Every item reports the same limitation: `run_batch` exposes total edge and retail/arb Y volume, but not the edge split. Volume deltas are given in the items in their place. Within the search items, j136's candidate trades less volume than T3: about −38 to −44 Y retail and −11 to −12 Y arb per market. j140's figures are about −76 to −78 retail and −23 to −25 arb. For replica deltas over T3 on the holdout, j140 reports retail −0.0232 and arb +0.0460. - **Gate C reporting is inconsistent, but it does not matter here.** j133 and j134 mark gate C as "passed" in real-only mode by reusing gate A on the unchanged T3. j131, j132, j135, j137, j138 and j139 mark it not applicable. Only the two search items froze new Solidity, and both passed gate C (j136 6.29e-05 bps, j140 5.71e-05 bps, threshold 0.1). - **j136 records less gate detail than j140.** It omits the gate A seed range and threshold and does not describe its freeze procedure. The item says only that *"Only the named constants and 19-8-2 head changed in Solidity"*. Its gate A max of 5.747e-05 bps matches every other item. ## Agreement across items - All 10 items report the same gate A maximum, 5.7471258602809216e-05 bps (50 sequences, 100,000 callbacks). All 10 also report the same gate B, CG−T = +1.920372244773669 se 0.04446, inside 2.0 ± 0.15. - j134 ran a holdout on T3 over T and got +2.533096306362695 se 0.0135. This is identical to j140's `base_over_T` on the same seeds 40000000..40008191. j136's over_T minus over_base, 2.5621 − 0.0290, also comes to 2.5331. - The start-of-search monitor reads +2.5846 in both j136 and j140. - The T3−HA blocks span +1.403 to +1.454. The largest gap between any two blocks (0.051) is within about 1.3 se of their difference, so the blocks do not disagree. ## Sources No item file contains a source list. j131's notes say its sources.json was empty, and j132 says its artifact references are in a sources.json that this merge did not receive. The source flags for j132–j140 therefore cannot be matched to specific claims here. Everything above rests on local measurements recorded in the item results, plus hash checks repeated in this merge, not on quoted web pages. The claims that could depend on external artifacts are the input-artifact hash confirmations, such as T3 2333ecf2…, CG ebbd90c6…, replica e8a7a2af… and HA 2801efc8…. That is especially true for j134, which has 4 flagged sources. These claims are consistent across all 10 items, but this merge did not re-download those artifacts. ## Merged data (JSON; numbers as strings) Under `best`, the `params` object is copied whole from j136 with its original numeric types, as the brief asks for params to be copied whole. `best.code_zb64` is j136's own string, and it decodes to `best.code_sha256` (checked here). ```json { "rows": [ { "key": "j131", "mode": "real-only", "result_sha256": "ae794344b7ecf827589ba4e6186a30f23d94e80f33d1ab4624f4b4930a5983cd", "markets_per_s": "172.71", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "237600", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j132", "mode": "real-only", "result_sha256": "e396d5b6e5cdf811650feee40334521814affe8999e717ac7bcc47c3cc2b2a73", "markets_per_s": "227.89282332476665", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "237645", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j133", "mode": "real-only", "result_sha256": "af5993718777e246342daf0ac283f2775377b0a6f97636b068b059cf51200d28", "markets_per_s": "173.30168165180083", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "237587", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j134", "mode": "real-only", "result_sha256": "b078649aff6e1227a6c8777b8b6b4bb596814ce4af6c6e21fd5bf07848252814", "markets_per_s": "252.8779648330379", "generations": null, "search_minutes": null, "holdout_over_T": { "mean": "2.533096306362695", "se": "0.01349951591946062" }, "holdout_over_base": { "mean": "0", "se": "0" }, "real_candidate_minus_T3": null, "afterSwap_peak": "237642", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j135", "mode": "real-only", "result_sha256": "e96e0a38d9c1b2dcd518da54a376d2be92a1e6347a84bfa09b020558010aaa86", "markets_per_s": "472.78285596779057", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "232525", "OOG": "0", "valid": "false", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j136", "mode": "search", "result_sha256": "4a7cb46ef64b112057a844a3da49912ce31e079134047feec41f0615157de67d", "markets_per_s": "1103.2164603252288", "generations": "287", "search_minutes": "56.27218445451775", "holdout_over_T": { "mean": "2.5620683461263423", "se": "0.013634123833516443" }, "holdout_over_base": { "mean": "0.028972039763647442", "se": "0.004298167204488982" }, "real_candidate_minus_T3": [ { "seeds": [ "9240000", "9241023" ], "n": "1024", "mean": "0.05448739774964673", "se": "0.012324113487944365", "sum": "55.79509529563825", "sumsq": "162.1462799998221" }, { "seeds": [ "9242000", "9243023" ], "n": "1024", "mean": "0.024012190589986215", "se": "0.014210696370648215", "sum": "24.588483164145885", "sumsq": "212.13715060375097" } ], "afterSwap_peak": "237706", "OOG": "0", "valid": "true", "last_curve_point": [ "55.455071903516846", "2.6058069655945975", "0.19126213658867858" ], "code_sha256": "f65ab9b7d79a3bc626b685da15847db13f804b57338ec3ca6742ca5315e96fb2" }, { "key": "j137", "mode": "real-only", "result_sha256": "96cac62db3b2ab88ced9457ed7165ec8b8a7df08d18b153dccd4cdfec9a4aac0", "markets_per_s": "426.51", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "237512", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j138", "mode": "real-only", "result_sha256": "5ca3ee1511e269e4a87fbaf57f6bca8031662345f285415466ab00c2f8943730", "markets_per_s": "214.0", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "237600", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j139", "mode": "real-only", "result_sha256": "ccd937a662d6e6329219123b35e44e46e16cc3e6e90bc0d2af994788f587895a", "markets_per_s": "202.1616690068629", "generations": null, "search_minutes": null, "holdout_over_T": null, "holdout_over_base": null, "real_candidate_minus_T3": null, "afterSwap_peak": "237461", "OOG": "0", "valid": "true", "last_curve_point": null, "code_sha256": "cb6d945511e1e10af78baa64c2a6cb25806f0998403ce6440654bfc3c70092d9" }, { "key": "j140", "mode": "search", "result_sha256": "deac4e1e668653a6e081adc85d7016aaed24b0b395fd6d307703d584e365f9b7", "markets_per_s": "957.7", "generations": "168", "search_minutes": "45.02", "holdout_over_T": { "mean": "2.555919598409799", "se": "0.013509966940197382" }, "holdout_over_base": { "mean": "0.022823292047104204", "se": "0.003730699926503159" }, "real_candidate_minus_T3": [ { "seeds": [ "9256000", "9257023" ], "n": "1024", "mean": "0.02104159213357215", "se": "0.010414652350301562", "sum": "21.546590344777883", "sumsq": "114.07608504262987" }, { "seeds": [ "9258000", "9259023" ], "n": "1024", "mean": "0.0294868861427367", "se": "0.009841158698279439", "sum": "30.194571410162382", "sumsq": "102.34408374598078" } ], "afterSwap_peak": "237771", "OOG": "0", "valid": "true", "last_curve_point": [ "45.05", "2.6115", "0.1375" ], "code_sha256": "807b2175998dd629761f78ff9f817cb20d0a7af60ffd1d2af5f9e8bea94ed93e" } ], "best": { "key": "j136", "holdout_over_T": { "mean": "2.5620683461263423", "se": "0.013634123833516443" }, "code_sha256": "f65ab9b7d79a3bc626b685da15847db13f804b57338ec3ca6742ca5315e96fb2", "code_zb64": "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", "params": { "W1": [ [ -0.165189, 0.710229, -0.315169, 3.132864, -2.507105, 0.728151, -1.574226, 2.449611, -5.24252, -1.718123, -0.746289, -0.029437, 0.467645, -2.131552, -0.595491, -0.265657, -1.230576, -0.254637, 1.511416 ], [ 0.808829, 0.87899, 2.03786, 2.742503, 0.042498, -0.329561, -0.366972, 0.421396, -0.388023, 0.177141, -0.093909, -1.9214, 0.119695, -0.092553, -0.125954, 0.955415, 0.297363, 0.000706, 0.957639 ], [ 1.09084, 3.74764, -1.205413, 0.757015, 0.915537, 2.093602, -1.728015, 0.900777, 1.02407, -0.530529, 0.311215, 0.003682, -0.165924, 0.462082, 0.192026, 1.181019, -2.461336, -0.410413, -2.529686 ], [ 0.296841, -1.32655, 0.855797, -0.237493, -0.947823, -0.406769, 0.185772, 0.076605, 0.812665, -0.034584, 0.190298, -1.707974, -0.180372, 0.289318, 0.241528, -1.421208, 1.04605, -0.247339, -0.738774 ], [ 1.463351, 0.448487, 4.966308, 0.105702, -2.029062, -1.488895, -1.184732, -0.593242, -0.233336, 0.355127, -0.041532, -2.406298, -0.118469, -1.196982, 0.298133, -0.80861, 2.283354, 0.055208, 0.419967 ], [ 1.194508, -1.850602, 0.40392, -1.290774, 3.768444, 1.597409, 3.775852, 0.49459, -0.478043, 0.015736, 0.037778, 3.696492, 0.738403, -1.143243, 0.180776, 0.376947, -1.848528, -0.255722, -1.501154 ], [ -1.119841, 1.36318, 0.052747, -0.744233, 2.433904, 0.682839, -1.322077, 2.299101, 1.13886, -1.459251, 1.305436, 1.122782, 1.225612, -1.058731, -1.175275, -1.709284, 1.04665, -0.871725, 2.63061 ], [ -2.712548, -2.577082, 3.233453, 3.649093, -0.156138, -0.886509, -1.65865, 2.490426, 0.363074, -0.885558, 2.519369, -2.162098, 1.295385, -0.020157, 0.546411, 0.944699, -0.471837, -1.705973, 1.123898 ] ], "b1": [ 2.109718, -1.815726, -4.320571, -0.804923, -1.530634, -2.252836, -0.594247, 0.593669 ], "W2": [ [ 0.144377, -2.071611, 0.942834, -0.586287, -3.135636, -0.765434, -0.972671, 1.037108 ], [ 0.427327, -2.068181, -1.77965, 3.508785, 0.324103, -2.152281, 0.774569, -0.218668 ] ], "b2": [ 1.550478, 2.838522 ], "constants": { "ELAPSED_CAP": 7, "SIGNAL_THRESHOLD": 0.001175371984665033, "DIR_DECAY": 0.7402601547682512, "ACT_DECAY": 0.38358832908630014, "SIZE_DECAY": 0.47259515144757647, "TOX_DECAY": 0.9271021277435694, "SIGMA_DECAY": 0.7336343026525487, "LAMBDA_DECAY": 0.9919320806253972, "SIZE_BLEND_DECAY": 0.8382585245088601, "TOX_BLEND_DECAY": 0.2312683735106681, "ACT_BLEND_DECAY": 0.9874508071382325, "PHAT_ALPHA": 0.2511717177720851, "PHAT_ALPHA_RETAIL": 0.05422107894132092, "PHAT_ALPHA_SAME": 0.14500316310613717, "DIR_IMPACT_MULT": 4, "GATE_SIGMA_MULT": 6.5616864249544715, "MIN_GATE": 0.02937313062739035, "TRADE_TOX_BOOST": 0.3061065680818132, "STALE_ATTRACT_FRAC": 0.6740388135583445, "RET_CAP": 0.02249981403583355, "TOX_CAP": 0.2059650095145359, "TRADE_RATIO_CAP": 0.17939137928076143, "LAMBDA_CAP": 1.0813179608505465, "STEP_COUNT_CAP": 66, "BASE_FEE": 0.000407595737398728, "SIGMA_COEF": 0.20074675383568513, "LAMBDA_COEF": 0.001334383134674663, "FLOW_SIZE_COEF": 0.4949741506166612, "TOX_COEF": 0.030029413465258906, "TOX_QUAD_COEF": 1.58373047577747, "ACT_COEF": 8.600717833960108, "DIR_COEF": 0.001778444246064178, "DIR_TOX_COEF": 0.00776721587465968, "STALE_DIR_COEF": 0.7503014752670941, "SIGMA_TOX_COEF": 0.05947126138660496, "TAIL_KNEE": 0.05494672565162884, "TAIL_SLOPE_PROTECT": 0.8618569792386532, "TAIL_SLOPE_ATTRACT": 0.8519043387584746, "ANCHOR_COUNT_CAP": 710, "ANCHOR_LINEAR_K": 0.6531346776657709, "ANCHOR_QUADRATIC_K": 1.0944325159315949 }, "scaling": { "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_size", "step_trade_count", "prev_bid_fee", "prev_ask_fee", "steps_since_arb_like_print", "router_implied_vanilla_spot_dev" ], "feature_multipliers": [ 1000.0, 2.0, 500.0, 500.0, 500.0, 1.0, 500.0, 500.0, 2.0, 2.0, 1000000.0, 1.0, 500.0, 647.054817, 1.395287, 220.617911, 228.148131, 0.412565, 274.335453 ], "feature_offsets": [ -1.0, -1.0, -1.0, -1.0, -1.0, -1.0, 0.0, 0.0, -1.0, -1.0, -1.0, 0.0, -1.0, -0.618293, -2.037262, -0.803057, -0.808615, -0.823811, 0.001772 ], "weight_fixed_point_scale": 1000000, "input_clip": [ -4, 4 ], "hidden_relu_clip": [ 0, 4 ], "output_unit_fee": 0.0001, "note": "unchanged from CG" } } }, "real_only": [ { "key": "j131", "blocks_T3_minus_HA": [ { "seeds": [ "9220000", "9220447" ], "n": "448", "mean": "1.4414966760718697", "se": "0.048347304067281396", "sum": "645.7905108801976", "sumsq": "1398.995627210899" } ], "gas": { "T3_afterSwap_peak": "237600", "T3_OOG": "0", "HA_afterSwap_peak": "231567", "HA_OOG": "0" } }, { "key": "j132", "blocks_T3_minus_HA": [ { "seeds": [ "9224000", "9225023" ], "n": "1024", "mean": "1.4527051517754659", "se": "0.029938525941747125", "sum": "1487.570075418077", "sumsq": "3099.9376345882956" } ], "gas": { "T3_afterSwap_peak": "237645", "T3_OOG": "0", "HA_afterSwap_peak": "233576", "HA_OOG": "0" } }, { "key": "j133", "blocks_T3_minus_HA": [ { "seeds": [ "9228000", "9229023" ], "n": "1024", "mean": "1.4279027539341238", "se": "0.03015635332912321", "sum": "1462.1724200285428", "sumsq": "3040.489728680066" } ], "gas": { "T3_afterSwap_peak": "237587", "T3_OOG": "0", "HA_afterSwap_peak": "233659", "HA_OOG": "0" } }, { "key": "j134", "blocks_T3_minus_HA": [ { "seeds": [ "9232000", "9233023" ], "n": "1024", "mean": "1.4205904707768104", "se": "0.02680828136458995", "sum": "1454.684642075454", "sumsq": "2819.369949418015" }, { "seeds": [ "9234000", "9235023" ], "n": "1024", "mean": "1.409415917300676", "se": "0.02813116167173564", "sum": "1443.2418993158922", "sumsq": "2863.121220457333" } ], "gas": { "T3_afterSwap_peak": "237642", "T3_OOG": "0", "HA_afterSwap_peak": "233693", "HA_OOG": "0" } }, { "key": "j135", "blocks_T3_minus_HA": [], "gas": { "T3_afterSwap_peak": "232525", "T3_OOG": "0", "HA_afterSwap_peak": null, "HA_OOG": null } }, { "key": "j137", "blocks_T3_minus_HA": [ { "seeds": [ "9244000", "9245023" ], "n": "1024", "mean": "1.4031391548837386", "se": "0.02686964914195536", "sum": "1436.8144946009484", "sumsq": "2772.360220687264" } ], "gas": { "T3_afterSwap_peak": "237512", "T3_OOG": "0", "HA_afterSwap_peak": "233591", "HA_OOG": "0" } }, { "key": "j138", "blocks_T3_minus_HA": [ { "seeds": [ "9248000", "9249023" ], "n": "1024", "mean": "1.4540682149844733", "se": "0.02763873382506366", "sum": "1488.9658521441006", "sumsq": "2965.2824803863177" } ], "gas": { "T3_afterSwap_peak": "237600", "T3_OOG": "0", "HA_afterSwap_peak": "231592", "HA_OOG": "0" } }, { "key": "j139", "blocks_T3_minus_HA": [ { "seeds": [ "9252000", "9253023" ], "n": "1024", "mean": "1.4201825474911165", "se": "0.030471202926308755", "sum": "1454.2669286309033", "sumsq": "3037.9704757795826" } ], "gas": { "T3_afterSwap_peak": "237461", "T3_OOG": "0", "HA_afterSwap_peak": "231504", "HA_OOG": "0" } } ], "real_only_pooled_T3_minus_HA": { "n": "7616", "mean": "1.4277188580743325", "se": "0.010564808949150564", "sum": "10873.506823094116", "sumsq": "21997.52733720777" }, "search_pooled_candidate_minus_T3": { "j136": { "n": "2048", "mean": "0.03924979416981647", "se": "0.009408885691180232", "sum": "80.38357845978413", "sumsq": "374.28343060357304" }, "j140": { "n": "2048", "mean": "0.025264239138154426", "se": "0.007163238289040846", "sum": "51.741161754940265", "sumsq": "216.42016878861065" } } } ```