Add Lost Cities selectivity research notes (4-model analysis)

Mechanism-level analysis of the slot_aware_playability iter 240
plateau (opened_colors 4.95+, bad_open_rate 88-92%, calibration
gap 6-9 → 2-4). Captures four expert consultations with diagnostic
hypotheses, intervention catalog (architectural / training-dynamics
/ game-specific), measurement plan, and a comparison table across
the four sources.

Key new directions surfaced:
- Current vs average vs league policy separation (Deep CFR average
  strategy is the convergence target, not advantage current).
- All-negative fallback as Deep CFR ablation lever.
- Empirical r̃ partitioning by action class.
- Tabular Lost Cities oracle as a clean test of "is 5-color the
  game-theoretic answer or an approximation artifact".
- Entry-gate target defined from traversal counterfactual values
  instead of heuristic labels.

ideas.md gets an "Active Research Threads" pointer.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

This commit is contained in:

coolguy

2026-05-07 20:58:09 +09:00

co-authored by

Claude Opus 4.7

parent 6c9babe769

commit 8f6780dd3d

2 changed files with 205 additions and 0 deletions

									
										ideas.md
									
		+4
		
												View File
												
				@@ -1,5 +1,9 @@

				# Lost Cities Deep CFR — Ideas 압축 정리

				## Active Research Threads

				- [Lost Cities selectivity 가설 / 4-model 분석](docs/research/lost_cities_selectivity.md)

				## 진단 가설들

				- Variance 가설 — Draw variance가 entry regret signal 압도. Reward smoothing 1순위.