id: c05233e1ddd74e89b550a3a4b45c4a70
parent_id: 5bf5676aed9e439f9934f2d603144744
item_type: 1
item_id: 47031623e2c6451fb36730ffe93380ea
item_updated_time: 1783540388132
title_diff: "[]"
body_diff: "[{\"diffs\":[[0,\"026-07-0\"],[-1,\"6\"],[1,\"8\"],[0,\".\\\n\\\n---\\\n\\\n\"]],\"start1\":22,\"start2\":22,\"length1\":17,\"length2\":17},{\"diffs\":[[0,\"Gen 2-4.\"],[-1,\"7\"],[1,\"8\"],[0,\" Strateg\"]],\"start1\":589,\"start2\":589,\"length1\":17,\"length2\":17},{\"diffs\":[[0,\"lop\\\n- ✅ \"],[-1,\"**\"],[0,\"Gen 4.7:\"]],\"start1\":928,\"start2\":928,\"length1\":18,\"length2\":16},{\"diffs\":[[0,\"EM2)\"],[-1,\" — CURRENT PRODUCTION**\\\n  - PFR-based raiser range with bet-sizing-aware narrowing\\\n  - Hard equity floor for 3-bets (min_reraise_equity=0.42)\\\n  - Milder looseness on reraise (looseness^0.3)\\\n  - Ppot-based call decisions\\\n  - Short-stack strategy removed (folds defer to G47)\\\n  - Tuned via 2-round multi-\"],[1,\"\\\n- ✅ **Gen 4.8: Formula-based postflop with observer ranges — CURRENT PRODUCTION POSTFLOP**\\\n  - Unified score: `value_score + ppot_w×ppot - npot_w×npot`\\\n  - `value_score = hs×(1 - npot_w×npot) + nutredraw_w×nutpot×(1-hs)`\\\n  - call_threshold with multiway pot-odds cap: `0.28/√opp` (non-river), `0.15/√opp` (river)\\\n  - Bluff gating on weak opponents, stack-off AllIn guard, position-aware sizing\\\n  - Observer-based range equity (shared with G45 via `build_ranges_and_equity()`)\\\n  - Cold-start priors: VPIP 0.30, PFR 0.15, AF 1.0, fold_to_bet 0.50\\\n  - MC budget: 40,000 samples (divided by opponent count, floor 8K)\\\n  - **Config: `cash_nl_g48_live.toml`**\\\n\\\n### G47 Preflop — CURRENT PRODUCTION PREFLOP\\\n- PEM2 matrix equity (zero MC — pure lookup)\\\n- Uni\"],[0,\"fie\"],[-1,\"l\"],[0,\"d s\"],[-1,\"weep (10+8 params)\\\n  - **Performance: +10.6 BB/hand vs G45 field**\\\n\\\n### Engine Fixes (rusty-marvin)\\\n- ✅ Crash fix: seat restoration + chip refund on `run_to_completion()` error\\\n- ✅ `NotEnoughPlayers` now logs warning (was silent)\\\n- ✅ `NoLimit { max_raises: Option<usize> }` field added\\\n\\\n### Open Tuning Items\"],[1,\"core: `equity + ppot_w×ppot + nutpot_w×nutpot - npot_w×npot`\\\n- Position-based thresholds, stack-depth sigmoid\\\n- **Config: `cash_nl_g47_live.toml`**\\\n\\\n### Recent Bug Fixes (2026-07-08)\\\n- ✅ Fixed `remaining_stack_share` double-counting bug (formula_postflop:887)\\\n- ✅ Relaxed multiway call margin 0.35→0.28 for non-river streets\\\n- ✅ Postflop fold logging upgraded to info level\\\n- ✅ PlayerId seat mismatch fix (bot_id tracking through seat changes)\\\n- ✅ Call amount reconciliation (downward + upward current_bet correction)\\\n- ✅ Blinds synthesis for missing blind-posting events\\\n\\\n### Live Performance (2026-07-08 session)\\\n- **~65 hands at Ballsy II (12.5K/25K)**\\\n- Big wins: T4 two pair +21.2M, A9 trips +12.97M, QTo two pair +6.49M, QQ overpair +1.84M\\\n- **~42.5M total winnings** from value hands\\\n- Concern: ATo called 4.5M (187BB) preflop facing multiple all-ins — G47 threshold too low for huge bets\\\n\\\n### Open Tuning Items\\\n- [ ] **G47 preflop: tighten threshold for very large bets** (ATo 187BB call issue)\"],[0,\"\\\n- [\"]],\"start1\":975,\"start2\":975,\"length1\":625,\"length2\":1766},{\"diffs\":[[0,\"xtures)\\\n\"],[1,\"- [ ] Per-player looseness (use individual VPIP instead of table VPIP)\\\n\"],[0,\"- [ ] Ch\"]],\"start1\":2878,\"start2\":2878,\"length1\":16,\"length2\":87},{\"diffs\":[[0,\"scraper \"],[1,\"v0.1.10 \"],[0,\"— DOM pa\"]],\"start1\":3226,\"start2\":3226,\"length1\":16,\"length2\":24},{\"diffs\":[[0,\" bot\"],[-1,\" switched to G47** (2026-07-06)\"],[1,\": G48 live (primary) + G47 live (fallback)**\\\n- Per-hand event/decision logging in `/tmp/super_marvin_hands/`\"],[0,\"\\\n\\\n##\"]],\"start1\":3404,\"start2\":3404,\"length1\":39,\"length2\":116},{\"diffs\":[[0,\"ing \"],[-1,\"🔄 TRAINING IN PROGRESS\"],[1,\"⏸️ PAUSED\\\n\\\n> Gen5 data collection STOPPED — focus on G47/G48 optimization first.\"],[0,\"\\\n\\\n##\"]],\"start1\":3728,\"start2\":3728,\"length1\":31,\"length2\":88},{\"diffs\":[[0,\"e)\\\n\\\n\"],[-1,\"### Data Collection\\\n- **3.79M transitions** collected overnight 2026-07-05 (G4 teacher)\\\n- Total accumulated: ~10M+ transitions across multiple collections\\\n- DAgger round 1: ~1.43M transitions\\\n\\\n\"],[0,\"### \"]],\"start1\":4363,\"start2\":4363,\"length1\":201,\"length2\":8},{\"diffs\":[[0,\"xt Steps\"],[1,\" (when resumed)\"],[0,\"\\\n1. ⬜ Up\"]],\"start1\":4755,\"start2\":4755,\"length1\":16,\"length2\":31},{\"diffs\":[[0,\"use **G4\"],[-1,\"7\"],[1,\"8\"],[0,\" as teac\"]],\"start1\":4814,\"start2\":4814,\"length1\":17,\"length2\":17},{\"diffs\":[[0,\"er**\"],[-1,\" (currently G4)\"],[0,\"\\\n2. \"]],\"start1\":4832,\"start2\":4832,\"length1\":23,\"length2\":8},{\"diffs\":[[0,\" on \"],[-1,\"3.79M G47-collected\"],[1,\"new\"],[0,\" tra\"]],\"start1\":4851,\"start2\":4851,\"length1\":27,\"length2\":11},{\"diffs\":[[0,\"13 vs G4\"],[-1,\"7\"],[1,\"8\"],[0,\" in simu\"]],\"start1\":4886,\"start2\":4886,\"length1\":17,\"length2\":17},{\"diffs\":[[0,\"aus)\"],[-1,\"\\\n5. ⬜ Postflop equity matrix for better postflop features\"],[0,\"\\\n\\\n--\"]],\"start1\":4950,\"start2\":4950,\"length1\":65,\"length2\":8},{\"diffs\":[[0,\"026-07-0\"],[-1,\"6\"],[1,\"8\"],[0,\")\\\n\\\n| Met\"]],\"start1\":4978,\"start2\":4978,\"length1\":17,\"length2\":17},{\"diffs\":[[0,\"ion \"],[-1,\"bot\"],[1,\"preflop\"],[0,\" | **G\"],[-1,\"en 4.\"],[1,\"4\"],[0,\"7** (\"],[1,\"`\"],[0,\"g47_\"]],\"start1\":5068,\"start2\":5068,\"length1\":27,\"length2\":28},{\"diffs\":[[0,\"flop\"],[-1,\" + g4_adaptive_postflop) |\\\n| Live integration | Torn — functional |\\\n| G47 vs G45 field | **+10.6 BB/hand** (3 seeds × 10K hands) |\\\n| G47 vs Flock field | **+10.6 BB/hand** |\\\n| Gen 5 best model | v11, eval_acc=45.2%, +27.9 chips/hand |\\\n| Gen 5 data collected | 3.79M transitions (overnight G4 teacher) |\\\n| PEM2 matrix | 8.7 MB, 10 metrics × 8 bins × 169² pairs |\\\n| Regression tests | 28 hand replay tests (all pass) |\\\n\\\n---\\\n\\\n## Immediate Next Session Priorities\\\n\\\n1. **Train Gen5 v13** on the 3.79M overnight transitions with G47 preflop teacher\\\n2. **Update gen5_collect_full** to use G47 as teacher for future collection\\\n3. **Live monitoring** of G47 performance (watch for preflop leaks)\\\n4\"],[1,\"` — PEM2 matrix) |\\\n| Production postflop | **G48** (`g48_formula_postflop` — formula + observer ranges) |\\\n| Live integration | Torn — functional, userscript v0.1.10 |\\\n| MC budget (live) | 40,000 (divided by opponents, floor 8K) |\\\n| PEM2 matrix | 8.7 MB, 10 metrics × 8 bins × 169² pairs |\\\n| Regression tests | 28 hand replay tests + 477 lib tests (all pass) |\\\n| Gen 5 best model | v11, eval_acc=45.2%, +27.9 chips/hand |\\\n| Live session 2026-07-08 | ~65 hands, ~42.5M winnings |\\\n\\\n---\\\n\\\n## Immediate Next Session Priorities\\\n\\\n1. **Fix G47 preflop large-bet threshold** — ATo called 187BB with Ace-Ten offsuit; need tighter threshold for bets >50BB\\\n2. **Live monitoring** of G48 postflop — watch for fold/call patterns with medium hands\\\n3. **Per-player looseness** — use individual VPIP instead of table VPIP\\\n4. **Re-run sweeps** as profile data accumulates\\\n5\"],[0,\". **\"]],\"start1\":5106,\"start2\":5106,\"length1\":696,\"length2\":862},{\"diffs\":[[0,\"** —\"],[-1,\" consider\"],[0,\" pos\"]],\"start1\":5988,\"start2\":5988,\"length1\":17,\"length2\":8},{\"diffs\":[[0,\"flop\"],[-1,\"\\\n5. **Per-opponent PFR wiring** — use observed PFR per seat, not table average\"]],\"start1\":6031,\"start2\":6031,\"length1\":82,\"length2\":4}]"
metadata_diff: {"new":{},"deleted":[]}
encryption_cipher_text: 
encryption_applied: 0
updated_time: 2026-07-16T07:16:43.358Z
created_time: 2026-07-16T07:16:43.358Z
type_: 13