Personalized Learner
Evidence-aware weakness detection, ML offline evaluation, and solver-authoritative practice recommendations
The active solver curriculum currently spans 3 river boards across 3 reference strategy packs (~33 eligible postflop decisions). The system distinguishes observed drill performance from broad poker proficiency. It reports empirical observations honestly and prevents unvalidated ML models from promoting over transparent baselines. No synthetic scores are fabricated.
Operating on transparent empirical rules and curriculum diversity.
undefined solver-verified training points.
0 distinct decision scenarios encountered.
No validated model promoted yet.
Background Training & Evaluation Pipeline
Executes causal point-in-time feature extraction, group-held-out validation, and baseline comparison.