Experiment controls
The underlying synthetic path generator matches the current Options Research Lab Pro market-process parameters. All strategies see the identical path.
What is being tested?
The target line continues to compound even after an underfunded signal. No cash contributions, taxes, fund distributions, special crash/stick-around rules, or leverage-management rules are included in SIG-001 v0.1.
Why the implementation is deliberately limited
Jason Kelly's public materials describe the core quarterly target mechanism and an 80/20 starting split, but also state that the complete plan contains additional rules. Before calling a later study an “exact 3Sig” test, we should send our rule specification to Jason and ask him to confirm it.
Public references used for the mechanism: jasonkelly.com/books/3sig/ and jasonkelly.com/3sig/.
Target sweep
Run the experiment to compare 0% through 6% quarterly target lines.
| Target | Mean CAGR | vs 100% stock | vs 80/20 rebalance | Mean DD | Avg equity | Reserve depleted |
|---|
Mean CAGR by quarterly target
3% target detail
Benchmarks
| Benchmark | Mean CAGR | Mean drawdown | Avg equity |
|---|
The 80/20 rebalancing control is critical: if Signal only looks safer because it holds less stock, this helps expose that.