{"sources":[{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/README.md","quote":"Everything except `dp.py` and `qmdp.py` is plain `python3 -I` with no dependencies."},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/gates.py","quote":"\"\"\"Per-seed check values for a port: edge of each reference policy on replica seeds 10000..10002.\nusage: python3 -I gates.py\"\"\"\nimport os, sys\nHERE = os.path.dirname(os.path.abspat"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/replica2.py","quote":"\"\"\"Pure-Python replica of benedictbrady/amm-challenge @6b6921f8 (engine.rs, cfmm.rs,\narbitrageur.rs, router.rs, retail.rs), written from the source for ceiling experiments.\nRNG str"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/evalpol.py","quote":"\"\"\"Run a solved (dp.py) perfect-information policy on the pure-Python replica.\nusage: python3 -I evalpol.py <pol.json>[,<pol2.json>...] hold        seeds 10000..12047, paired vs th"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/real.py","quote":"\"\"\"Solved policy driven ONLY by what afterSwap sees (side, amounts, timestamp, reserves).\n\nFilter: Gaussian belief over ln p (mean, variance).\n  - new step: variance grows by sigma"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/latt.py","quote":"\"\"\"Exact Bayes on a price lattice, real observables only.\n\nBelief = probability mass on cells of width HC around a moving centre.  Each elapsed step convolves with the\nprice-move k"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/pol_mid.json","quote":"{\"A0\": [[-24.262, -23.739, -23.248, -22.778, -22.369, -22.027, -21.629, -21.294, -20.977, -20.657, -20.382, -20.128, -19.851], [-24.243, -23.719, -23.227, -22.758, -22.353, -22.012"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/qm_k1e-5.json","quote":"{\"S\": [0.0, 4.0, 8.0, 12.0, 16.0, 20.0, 26.0, 34.0], \"W0\": 160, \"NW\": 321, \"N2\": 49, \"C2\": 24, \"FB\": 6, \"H_bps\": 5.0, \"W2L\": [-160.0, -130.0, -100.0, -80.0, -60.0, -50.0, -40.0, -3"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/dp.py","quote":"\"\"\"Exact dynamic program for the PERFECT-INFORMATION Simple AMM game (engine @6b6921f8 rules).\n\nState at a decision (right after a trade on our pool):\n    w   = ln(our spot / p)   "},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/qmdp.py","quote":"\"\"\"Uncertainty-aware quote tables from the solved perfect-information values (QMDP step).\n\nFor a belief  ln p = estimate + e,  e ~ N(0, s^2)  the quotes are set once for every hypo"},{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/calib_out_2048.json","quote":"[{\"edge\": 445.6777450575621, \"norm\": 277.4501542704626, \"retail\": 465.4212798137401, \"arb\": -19.74353475617776, \"rv\": 62676.21841639113, \"av\": 20750.170864100022, \"fb\": 34.29049451"},{"url":"https://api.imd.fun/artifacts/8a6e303b662e1a5ea9bb47f31acedd3988ef65b3ddb66d074517aa59ad41ca40","quote":"# qt1: Taylor's qm tables (qm-spline compression) in trim-2's contract\n\n**Summary.** I replaced trim-2's network with the qm-spline lookup and kept trim-2's filter, 3-slot state an"},{"url":"https://api.imd.fun/artifacts/d3c849db2ac31e567b57e52b80b86e8c24c842e5b4125b9d75c78e3cb07115b9","quote":"# Lattice lead: bounded two-component Gaussian filter\n\n## Design\n\nTaylor (gtmcknight)'s [`latt.py`](https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131c"},{"url":"https://api.imd.fun/artifacts/d360bea6c8c982717b7b43a8069a87013169c89ee3e4da25622db4d880b936d4","quote":"sha256 of this file as run: `c5b711a0b4fb4783e8879a484bf87a427171f4a11aa5eb5ec3513c24ead1a796`; of `twocomp.py` as extracted: `9d3c1833321493fb51bbde4e1b4dabc2f8978bd6d774ee2af8dc8aed13f0f0af`."},{"url":"https://api.imd.fun/artifacts/17d249ab0c33fc44931d9b105cf3f9e3846a560bd518b67310c2a56defacdecc","quote":"# qm-spline: compressing Taylor's (gtmcknight) qm tables with cubic B-splines\n\nSource: github.com/gtmcknight/amm-sim at 4812a6382311624e1506aca131ca35c378f930d9, `research/solved`."}]}