fix 2nd train/inference mismatch: last-move feature lost in FEN round-trip

fix 2nd train/inference mismatch: last-move feature lost in FEN round-trip

fix: repetition feature lost in FEN round-trip (train/inference mismatch); greedy-health exp

fix: repetition feature lost in FEN round-trip (train/inference mismatch); greedy-health exp

fix: repetition feature lost in FEN round-trip (train/inference mismatch); greedy-health exp

revert logp_old change (ablation-falsified); keep only_player + kl estimator fixes

add logp_raw ablation switch

adjudicate claimed repetition/50-move draws on material

draw-penalty: remove guaranteed-0 repetition optimum

PPO correctness: logp_old from behaviour dist (temp+dirichlet), not raw policy

PPO fix: exclude opponent positions from buffer + staleness bound

spawn-based persistent self-play pool (fork/libgomp deadlock fix)

parallel self-play (--workers), conversion experiments, whitelist

tinychess: full research artifact - self-play RL, ablations, compute scaling, faithfulness, precision, plasticity

tinychess: self-play research substrate (phase 1-4)