When2Buy · Option POD3
0DTE Iron Condor Real-Price Backtest
Diagnostic backtest using historical Alpaca option bars for SPY + QQQ 0DTE iron condors. This page is intentionally blunt: the current real-price backtest is not launch-ready; it exposes sizing / payoff-model risk that synthetic tests hid.
Generated 2026-05-13 03:27 UTCPeriod 2024-02-01 → 2026-05-11Real option bars, no synthetic fallback
Full trades
1,119
Full win rate
85.7%
Full total P&L
-$11.61M
Full Sharpe
-0.62
OOS trades
874
OOS win rate
84.7%
OOS total P&L
-$9.82M
OOS max drawdown
-$17.11M
Equity curve
Monthly P&L
Original backtest figure
What was tested
- Underlying universe: SPY + QQQ
- Structure: 0DTE iron condor
- Pricing source: Alpaca historical option bars
- Entry/exit: conservative bid/ask real-price replay
- Fallbacks: no synthetic option-price fallback for missing data
- Transformer unavailable in this environment; run used HMM/calendar path only, with HMM dependency missing so regimes collapsed to bull-only.
Next fixes before production
- Normalize P&L by real defined risk / contract count.
- Add final expiry settlement reconciliation.
- Restore HMM + transformer dependencies for regime filtering.
- Add mega-cap earnings / NVDA spillover filter.
- Re-run OOS after correcting payoff scaling.
Walk-forward folds
| Fold | Period | Trades | Win rate | P&L |
|---|---|---|---|---|
| F1 | 2024-08-01 → 2024-09-30 | 84 | 81.0% | -$2.90M |
| F2 | 2024-10-01 → 2024-11-29 | 86 | 89.5% | $2.00M |
| F3 | 2024-12-02 → 2025-01-31 | 79 | 87.3% | -$1.01M |
| F4 | 2025-02-03 → 2025-03-31 | 80 | 71.2% | -$3.94M |
| F5 | 2025-04-01 → 2025-05-30 | 84 | 70.2% | -$5.57M |
| F6 | 2025-06-02 → 2025-07-31 | 83 | 97.6% | $1.92M |
| F7 | 2025-08-01 → 2025-09-30 | 82 | 95.1% | $1.90M |
| F8 | 2025-10-01 → 2025-11-28 | 81 | 86.4% | -$2.03M |
| F9 | 2025-12-01 → 2026-01-30 | 78 | 92.3% | $1.50M |
| F10 | 2026-02-02 → 2026-03-31 | 81 | 75.3% | -$2.35M |
| F11 | 2026-04-01 → 2026-05-11 | 56 | 85.7% | $649K |