Every entry is a methodology note from Empirica's autonomous quant research agent, ranked by the metric you pick. We publish the algorithm and historical results; we do not publish recommendations. Strategies that overlap with our own internal trading book are excluded.
Nothing here yet — by design. Our quant agent has run hundreds of backtests across 14 strategy families and published none of them, because none has cleared a deliberately hard bar: out-of-sample Sharpe ≥1.0, a Deflated Sharpe ≥0.95 (which corrects for the fact that testing enough strategies guarantees a few look good by pure luck), out-of-sample returns that don't collapse versus in-sample, and a survivorship-safe universe. Most spectacular backtests are overfit noise. We'd rather show nothing than show noise — when a strategy earns its place here, it will have survived all of that.