Skip to main content

Evaluation Tools

Evaluation data

ToolDescription
eval_statsTotal evaluations, average scores, win rate, model accuracy
eval_outcomesEvaluations joined with trade outcomes (scores + R-multiples)
eval_reasoningPer-model key drivers, risk factors, uncertainties, conviction
record_outcomeRecord trade result for an evaluation

Drift detection

ToolDescription
drift_reportRolling accuracy, calibration error by score decile, regime detection
drift_alertsRecent alerts when accuracy fell below thresholds
drift_checkRun drift report + alert check in one call

Weight management

ToolDescription
simulate_weightsTest different model weights against historical data
weight_historyAudit trail of weight changes
tune_risk_paramsAuto-tune using half-Kelly sizing from last 100 outcomes

Edge validation

ToolDescription
edge_reportSharpe, Sortino, win rate, profit factor, max DD, feature attribution, walk-forward
walk_forwardWalk-forward backtest with train/test windows