Frequently Asked Questions

Sigmentic runs seven independent statistical tests on your trading strategy to identify overfitting, spurious correlations, regime fragility, and execution risk. Upload your backtest, get a clear verdict in under five minutes. The result is a Validation Passport you can share with investors or use to inform your own analysis.

An estimated 87% of backtested strategies fail in live trading. Traditional backtesting answers the wrong question: "would this have been profitable?" instead of "is this edge real?" Academic research shows most published trading factors don't replicate, and returns decay significantly after publication. Common failure modes include overfitting to noise, spurious correlations, regime dependency, and unrealistic execution assumptions.

The Validation Passport is what you receive after validation. It contains layer scores, a composite verdict, and the statistical evidence behind each result. The Passport is cryptographically signed and portable. Share it with investors, prop desks, or compliance teams without revealing your strategy logic.

Most validation tools test one thing: is this strategy overfit? That's necessary but incomplete. They don't test whether your edge is causal, whether it survives regime shifts, or how much your returns shrink after real execution costs. Sigmentic runs seven independent validation layers, each targeting a different failure mode. A strategy can pass standard validation and still fail on causality, regime robustness, or execution realism.

Pine Script (TradingView), Python, MQL5 (MetaTrader), natural language description, and direct API submission. All inputs are converted into a canonical format before flowing through each validation layer, so you can use whatever tools you already trade with.

Zero data retention on strategy inputs. Your logic is never stored or shared. For institutional users, Layer 7 offers confidential validation using trusted execution environments: your strategy runs inside a secure enclave, meaning no one (not even Sigmentic) can see your proprietary logic.

VALIDATED (80-100): strategy passes all critical statistical thresholds. CONDITIONAL (60-79): passes most tests, identified weaknesses require review. WEAK (40-59): significant statistical concerns detected. REJECTED (0-39): fails critical validation thresholds across multiple layers. These labels describe what our tests found, not what you should do with your capital.