Picks track record

← Back to picks

Cumulative performance of the daily-picks strategy vs an equal-weight all-universe benchmark. Each cohort = top-10 picks held for 20 trading days. Stats use non-overlapping 20-day cohorts to avoid double-counting overlapping holds.

Cumulative
+60.9%
vs +43.7% benchmark
Alpha per cohort
+0.55%
per 20-day rebalance
Sharpe
0.98
vs 1.37 benchmark
Max drawdown
-19.6%
of picks equity

Equity curve

Hypothetical $1 start, rebalanced every 20 trading days into the day's top-10 picks. Benchmark = equal-weight all-universe over the same 20-day windows.

Detail

Cohorts resolved
606 / 627
Cohort hit rate
54.8%
Avg picks 20d return
+1.79%
Avg benchmark 20d return
+1.23%

Caveats — read before you trust this

  • This is a BACKTEST on 2024-2026 data. The model was trained on pre-2024 and applied forward. Forward-looking data was not used in features.
  • Period is a strong bull market (S&P 500 ~+50% over the window). Strategy returns include market beta, not pure alpha. Performance in a bear market is unverified.
  • No transaction costs, slippage, or taxes assumed. Real performance would be lower by 1-3% annualized.
  • The model's edge comes mostly from macro-regime signals (DGS10, VIX, yield curve). It is NOT a per-stock alpha. Use as decision aid, not auto-trade signal.