{"item":"sigma-1","batch":"b10-tell-sigma","track":"B","status":"complete","verdict":"No incremental gain. Frozen SigmaTell vs identical TellGauss on holdout 10000..12047: +0.0019 se 0.0035 (n 2048). An oracle given each run's TRUE sigma (diagnostic, not blind) gains only +0.0055 se 0.0026 on 4096 seeds, so per-run sigma is worth at most ~0.01 even with near-exact tell pins; the earlier +0.022 pre-tell figure is not improved upon. Not worth porting to Track A.","mechanism":{"summary":"Causal per-run diffusion estimate from tell-inferred pins, used only in the two variance-growth lines (v += sigma^2) of the first-trade prediction and the silent-step truncation. Everything else in TellGauss is unchanged (C=1, soft=1, qm_k1e-5.json, likelihood r, soft posterior; no hard pin on a tell).","estimator":"sigma2_hat = (K0*SIG2 + sum_i min(D_i^2, zcap^2*k_i*(10.08e-4)^2)) / (K0 + sum_i k_i), bounded to [(8.82e-4)^2, (10.08e-4)^2]. D_i is the ln-price change between consecutive pins (a pin is a first trade of a step whose soft posterior pi >= pmin; its price is the arb band-edge reading lp_arb). k_i is elapsed steps between them, including all silent no-arb steps. The first anchor is ln 100 at t=-1.","selection_bias_handling":"Pin times are stopping times: the arb fires when the price leaves the band of the quotes in force, and the pin decision uses only information at that call. By Wald's second identity, E[sum D_i^2 - sigma^2 sum k_i] = 0, so the pooled ratio over all elapsed steps is consistent. Averaging D_i^2/k_i per pair, or keeping only short gaps, would be biased. Silent no-arb intervals enter through k_i. Measured bias on 4096 diagnostic runs: -0.0011 bps (final estimate).","contamination_control":"Pin gate pi >= pmin uses the soft tell posterior. False pins are also capped by the per-increment D^2 cap (zcap sd at the top of the model range).","shrinkage":"K0 pseudo-steps at 9.45 bps (SIG2 = 0.000945^2).","fixed_sigma_mode":"est=0 gives exactly TellGauss; checked bit for bit on every callback (fees, m, v)."},"frozen_settings":{"est":1,"K0":2000,"pmin":0.999,"zcap":6,"C":1,"soft":1,"qm":"qm_k1e-5.json","frozen_at_utc":"2026-11-10T18:09:32.576933+00:00","sigpol_sha256":"13e35eb21bc818ae943e0e924bd2b559f7ec36006dd44b52b253fbecf2b479ab","tellpol_sha256":"6d65cd387bbf2eef8eb48a695cf7e977777080e81e550abb0fcd71209186f0ee","selection":"4 configurations tried on 67000..67255 (two settings varied: K0 in {500,2000,8000} at pmin 0.999, and pmin 0.9 at K0 2000; zcap fixed at 6). The pre-set default (a) had the highest tune mean. All four are within 1.2 se of zero."},"seed_coverage":{"tune":[67000,67255],"confirm":[67256,67511],"holdout":[10000,12047],"holdout_missing_candidate":[],"holdout_missing_base":[],"holdout_chunks":"8 chunks of 256 seeds, each arm one run.py call per chunk (about 4 to 5 s each, under 470 s), each seed scored once per arm","diagnostic_only":[68000,72095]},"tune_vs_base":{"a_K0_2000_pmin_0.999":{"paired":{"n":256,"mean_diff":0.003594167296012918,"se":0.00953620668232354,"sum":0.920106827779307,"sumsq":5.939820467197369,"worst5":[[67088,-0.409],[67228,-0.401],[67000,-0.386],[67145,-0.384],[67069,-0.332]]},"candidate":{"n_seeds":256,"mean_edge":532.8749705227566,"first_trades":1795841,"misread":1075,"misread_rate":0.000598605333100202,"price_samples":2790683,"price_rms_bps":7.1798847962386105}},"b_K0_500_pmin_0.999":{"paired":{"n":256,"mean_diff":-0.006748980673030092,"se":0.01033401522040496,"sum":-1.7277390522957035,"sumsq":6.98303378864463,"worst5":[[67061,-0.549],[67000,-0.461],[67148,-0.418],[67235,-0.407],[67145,-0.388]]},"candidate":{"n_seeds":256,"mean_edge":532.8646273747876,"first_trades":1795810,"misread":1203,"misread_rate":0.000669892694661462,"price_samples":2790577,"price_rms_bps":7.1908078005682645}},"c_K0_8000_pmin_0.999":{"paired":{"n":256,"mean_diff":-0.0070598268569590505,"se":0.008393881732828228,"sum":-1.807315675381517,"sumsq":4.61220865130256,"worst5":[[67239,-0.498],[67176,-0.443],[67129,-0.337],[67000,-0.292],[67202,-0.282]]},"candidate":{"n_seeds":256,"mean_edge":532.8643165286036,"first_trades":1795837,"misread":1181,"misread_rate":0.0006576320679438055,"price_samples":2790659,"price_rms_bps":7.189548448518744}},"d_K0_2000_pmin_0.9":{"paired":{"n":256,"mean_diff":-0.01221739634944674,"se":0.010733825543119043,"sum":-3.1276534654583656,"sumsq":7.559447686409921,"worst5":[[67202,-1.25],[67236,-0.564],[67034,-0.551],[67006,-0.474],[67018,-0.434]]},"candidate":{"n_seeds":256,"mean_edge":532.8591589591111,"first_trades":1795823,"misread":1184,"misread_rate":0.0006593077380120424,"price_samples":2790626,"price_rms_bps":7.18777272394394}}},"confirm_vs_base":{"paired":{"n":256,"mean_diff":0.015668760244853086,"se":0.009263945937317436,"sum":4.01120262268239,"sumsq":5.665225498020723,"worst5":[[67479,-0.404],[67407,-0.37],[67460,-0.367],[67343,-0.352],[67390,-0.336]]},"candidate":{"n_seeds":256,"mean_edge":536.8837908553816,"first_trades":1793845,"misread":1190,"misread_rate":0.0006633795004585123,"price_samples":2786802,"price_rms_bps":7.190890008705908},"reference":{"n_seeds":256,"mean_edge":536.8681220951368,"first_trades":1793827,"misread":1188,"misread_rate":0.0006622712223642525,"price_samples":2786792,"price_rms_bps":7.19665353255038}},"holdout_vs_base":{"paired":{"n":2048,"mean_diff":0.0018643164979594562,"se":0.0034988317346925164,"sum":3.8181201878209663,"sumsq":51.32797623551785,"worst5":[[11149,-1.619],[11488,-0.73],[11807,-0.678],[11130,-0.615],[11916,-0.597]]},"candidate":{"n_seeds":2048,"mean_edge":536.5213012688791,"first_trades":14351653,"misread":9280,"misread_rate":0.0006466154107822981,"price_samples":22284268,"price_rms_bps":7.203290039678291},"reference":{"n_seeds":2048,"mean_edge":536.5194369523812,"first_trades":14351283,"misread":9362,"misread_rate":0.0006523458564645405,"price_samples":22283481,"price_rms_bps":7.2067238720673865}},"holdout_vs_lattice_qm":{"paired":{"n":2048,"mean_diff":2.7003768947177256,"se":0.02086416094710407,"sum":5530.371880381902,"sumsq":16759.03286997796,"worst5":[[10403,-1.081],[10578,-0.963],[10923,-0.565],[10387,-0.331],[10731,0.003]]},"candidate":{"n_seeds":2048,"mean_edge":536.5213012688791,"first_trades":14351653,"misread":9280,"misread_rate":0.0006466154107822981,"price_samples":22284268,"price_rms_bps":7.203290039678291},"reference":{"n_seeds":2048,"mean_edge":533.8209243741613,"first_trades":14315685,"misread":2322695,"misread_rate":0.162248261260289,"price_samples":22263118,"price_rms_bps":13.64283973652866}},"holdout_vs_flag_qm":{"paired":{"n":2048,"mean_diff":-0.02155995565663371,"se":0.0033897018867492194,"sum":-44.15478918478584,"sumsq":49.1213274263284,"worst5":[[11149,-1.756],[11488,-0.726],[11014,-0.571],[11916,-0.564],[11795,-0.534]]},"candidate":{"n_seeds":2048,"mean_edge":536.5213012688791,"first_trades":14351653,"misread":9280,"misread_rate":0.0006466154107822981,"price_samples":22284268,"price_rms_bps":7.203290039678291},"reference":{"n_seeds":2048,"mean_edge":536.5428612245357,"first_trades":14349512,"misread":0,"misread_rate":0.0,"price_samples":22279454,"price_rms_bps":7.1119119607811285}},"holdout_base_vs_flag_qm_local":{"paired":{"n":2048,"mean_diff":-0.023424272154593168,"se":0.00280105877574458,"sum":-47.97290937260681,"sumsq":34.015878714956095,"worst5":[[11712,-1.596],[11358,-0.663],[10636,-0.656],[11010,-0.566],[11172,-0.552]]},"note":"Local TellGauss holdout reproduction. tg-2 reported -0.02576 vs flag. Locally -0.02342, base mean edge 536.5193 vs tg-2's 536.5171. Same source hashes; the difference is attributed to the Python build (3.15 here), not investigated."},"diagnostics":{"oracle_true_sigma_vs_base_68000_72095":{"paired":{"n":4096,"mean_diff":0.005453508041186911,"se":0.002594962754465089,"sum":22.337568936701587,"sumsq":113.0691852264254,"worst5":[[69231,-1.116],[68169,-1.031],[70732,-0.768],[68660,-0.761],[69367,-0.746]]},"candidate":{"n_seeds":4096,"mean_edge":533.0829109112414,"first_trades":28671115,"misread":18448,"misread_rate":0.0006434350390628338,"price_samples":44468234,"price_rms_bps":7.1979271805626315},"reference":{"n_seeds":4096,"mean_edge":533.0774574032001,"first_trades":28670901,"misread":18622,"misread_rate":0.0006495087126839857,"price_samples":44468386,"price_rms_bps":7.203213482641612}},"estimator_accuracy_68000_72095":{"n_runs":4096,"final_bias_bps":-0.0010736964401034305,"final_rms_bps":0.18234429113614462,"at_t2500_bias_bps":-0.001474784056626197,"at_t2500_rms_bps":0.2864815712797789,"fixed_9.45_rms_bps":0.3674087662163254,"pins_per_run":3429.598388671875},"answer_to_brief":"Nearly exact pins do make sigma estimable: final error 0.18 bps RMS, against 0.37 bps for the fixed prior. But even exact sigma is worth only +0.0055 se 0.0026, so the earlier +0.022 result does not grow. The filter's variance is reset to 0 at about 3,430 pins a run, so a \u00b17% sigma error barely moves the sd used for table lookup or the arb likelihood."},"misread_and_price":{"holdout_candidate":{"first_trades":14351653,"misread":9280,"misread_rate":0.0006466154107822981,"price_rms_bps":7.203290039678291},"holdout_base":{"first_trades":14351283,"misread":9362,"misread_rate":0.0006523458564645405,"price_rms_bps":7.2067238720673865}},"gas_state_cost":{"measured":false,"estimate":"State: 4 extra words (last pin ln-price, last pin time, sum D^2, sum k) plus the current sig2. Can be packed into 2 slots. Per first-trade call that is a pin: one subtract, square, cap compare, two adds, one division and two bound compares, about 1-2k gas mostly SSTORE of the updated slots (2 warm slot writes ~ 2.9k each if not already dirty). Non-pin calls: one SLOAD of sig2 (2.1k cold or 100 warm). The estimate is not from a Solidity build; no EVM port was made."},"checks":{"all_pass":true,"est0_equals_tellgauss_bitwise":true,"replay_twice_equal":true,"tape_replay_into_fresh_policy_equal":true,"hidden_input_scramble_equal":true,"frozen_differs_from_base":true,"seeds":[67000,67001,67002,67123,67255,67256],"callbacks":62750,"ast":{"on_trade_args":["self","is_buy","ax","ay","t","x","y"],"forbidden_names_used":[]},"package_reextracted_and_confirm_rerun_bitwise_identical":true,"harness":"unchanged run.py/pair.py from the pinned solved folder (cmp-identical)"},"limitations":["Replica only. No engine, Solidity, gas measurement or board entry.","Tuning searched 4 configurations. All were statistically null, so selection carries no information.","Oracle and estimator-accuracy diagnostics read the true sigma outside the policy, on seeds 68000..72095. They are not candidates.","The estimator ignores the -sigma^2/2 per-step drift of ln p, which is negligible (about 4.5e-7 a step), and the 0.99 x arb clamp of the replica.","Only the variance-growth use of sigma was tested, as the brief required. No order-rate, quote or table changes."],"code_zb64":"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","code_format":{"encoding":"base64(zlib(tar))","tar_sha256":"7b1510288c7dba81d87979eb3a2812133f2080acc25a167dfbc824e13c640be4","tar_bytes":81920,"files":["MANIFEST.sha256","README.sigma.txt","SOLVED_SHA256SUMS","checks.json","checks.py","diag.py","evalpol.py","freeze.json","pair.py","par.py","real.py","replica2.py","reproduce.sh","run.py","score.sh","sigpol.py","summarize2.py","tellpol.py"],"tables_included":false,"reproduce":"place pinned solved-folder DATA beside the files (see reproduce.sh header), then bash reproduce.sh tune|confirm|holdout|diag|checks|summary"}}