Screener
2569 stocks Β· As of 2026-05-01 01:00:00+01:00 Β· Cross-column filters + composite multi-criterion sort
Filters
%
$
Arrows in column names show direction of goodness: β higher is better, β lower is better. top % / bottom % ops keep stocks above the chosen percentile. Add multiple "extra" columns under the column picker to compose a multi-factor percentile filter.
Filter funnel
live count after each filter
Universe
β
β No stocks match all filters. Try relaxing the filter that dropped the count to 0.
Sort (multi-criterion composite β drag-equivalent: weights add to 1.0 internally)
Backtest configuration
Uses your current filters + sort. Technical features are point-in-time correct. Fundamental filters use current snapshot only (historical EDGAR XBRL ingest is on the roadmap β task #52). Subsequent backtests reuse the cached panel and return in seconds.
CAGR
Sharpe
Sortino
Calmar (CAGR/MDD)
Max Drawdown
Annual vol
PSR (Sharpe>0)
PSR (Sharpe>1)
Deflated Sharpe
Walk-forward consistency (5 folds)
Mean Sharpe:
Β±:
Min:
Max:
Folds >0:
| Fold | Start | End | N | Sharpe | Sortino | CAGR | MaxDD |
|---|---|---|---|---|---|---|---|
A robust strategy should have β₯80% of folds with positive Sharpe AND a tight Sharpe band. PSR>0 above 95% means the result is statistically significant given sample size + higher moments. Deflated Sharpe adjusts for multiple-testing inflation (assumes β20 trials).
π² Random-baseline test (Serious Backtester)
Percentile:
p-value:
Random p50:
Random p95:
Sharpe 95% CI:
200 Monte Carlo replications with the same N stocks / rebalance freq / costs but RANDOM stock selection. If actual percentile > 95% AND p < 0.05, the strategy beats random with statistical significance. The Sharpe 95% CI is from block-bootstrap (preserves autocorrelation).
Fama-French 5-factor decomposition
Ξ± (annual):
Ξ± t-stat:
Residual Sharpe:
RΒ²:
| MKT-RF (Ξ²) | SMB (size) | HML (value) | RMW (profit) | CMA (invest) |
|---|---|---|---|---|
Decomposes returns into real Ξ± vs known risk premia (Fama-French 5: market, small-cap, value, profit, investment). |t-stat|>1.96 means Ξ± is significantly >0 β real edge, not factor exposure. Residual Sharpe is the pure-Ξ± Sharpe after factor exposures are stripped out.
Total return ,
Win rate ,
Ulcer ,
N rebalances .