Hypothetical results only
Track record
What we tested, how we tested it, and what came out, including the claims that failed. Results come from a pre-registered, out-of-sample backtest and a forward ledger of published postures. Neither is a record of actual trading.
Pre-registered backtest
Evidence run pending — we publish results only after an out-of-sample, pre-registered test. Hypothetical results only.
When the run is done, this page will show every claim we tested, whether it passed or failed, the out-of-sample sample size and confidence interval, and the hash of the pre-registration that fixed the rules before any result was seen.
Forward ledger
Forward ledger not yet published. Each trading day the published posture will be hashed and chained before the 09:35 ET decision time, then graded after the close with fixed rules. Hypothetical results only.
How to read this page
- Pre-registered: parameters, pass/fail gates and the train/test split were fixed and hashed before the test ran.
- Out-of-sample: only sessions the model never saw while it was being built count toward a result.
- Hypothetical: fills, costs and slippage are modelled. No account traded these results.
Full risk disclosures: Risk Disclaimer.