{"key": "qm2h-1", "check": {"calls": 1000, "max_offset_diff_bps": 16.05, "passed": false}, "settings_tried": [{"setting": "exact: gam 1, step 0.2, pmin 0.001", "n": 256, "mean_diff": 0.2397, "se": 0.034}, {"setting": "sdlev 1 (choice interpolated between qm sd levels)", "n": 256, "mean_diff": 0.2757, "se": 0.0323}, {"setting": "gam 1.5 (pi^1.5 weight)", "n": 256, "mean_diff": 0.2142, "se": 0.0356}], "confirm_paired": {"n": 256, "mean_diff": 0.2542, "se": 0.0307}, "holdout_vs_gauss_qm": {"n": 2048, "mean_diff": 0.2418, "se": 0.011, "sum": 495.1849, "sumsq": 625.0302, "worst5": [[10942, -2.072], [11555, -1.975], [11137, -1.974], [10401, -1.598], [10922, -1.569]]}, "holdout_vs_lattice_qm": {"n": 2048, "mean_diff": -0.2531, "se": 0.0139, "sum": -518.3714, "sumsq": 938.294, "worst5": [[10942, -3.182], [10035, -2.623], [10009, -2.607], [10403, -2.406], [11119, -2.357]]}, "misread_rate": {"base": 0.1728, "mine": 0.1698}, "price_error_bps": {"base": 14.38, "mine": 14.12}, "gas_estimate": null, "code_zb64": "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", "code_sha256": "1623d9786a6704a7842486913d85130921430e417ab8b855e252cf7adeecafc3", "notes": "Exact two-reading expected-value chooser on first trades with 0.001<pi<0.999; every other call is real.py soft=1 qm=qm_k1e-5.json. Value arrays from dp.py DP_SAVEVN=1 (110 sweeps, 536.821, tables identical to pol_mid.json). Check: on 1000 recorded first-trade calls (seed 60000) the chooser at pi=1 and pi=0 matched the qm table lookup of each reading within 5 bps on 994 of 1000 calls (2000 comparisons: 6 over 5 bps, max 16.05, mean 1.24 arb / 1.36 retail); with sd-level interpolation 995/1000, max 10.3. The residual cases are near-tied two-mode action values at displaced w where the table interpolates offsets between rows/levels; so passed=false on the strict max criterion, and I measured anyway. Frozen setting: gam 1, step 0.2 lattice, pmin 0.001, sdlev 0. Confirm +0.254. Holdout +0.242 vs gauss_qm, -0.253 vs lattice_qm. Holdout ran in 256- and 128-seed commands (512 would exceed 8 min here); one 256 chunk timed out and was rerun as two 128 chunks, no seed filled. gas_estimate null: the chooser needs the 260 MB value arrays and a 2-D search per call; not portable as is. code_sha256 is sha256 of the tar before zlib."}