Backtest quality is often determined by the mistakes it prevents. Before trusting a performance number, verify the data, calculation and assumptions that created it.
Data and basket definition
Confirm the exact delivery month and year of every leg in every historical season. Check units, multipliers, weights and sign. Require synchronized prices, and report missing years rather than substituting zeros or stale values. Keep any roll methodology explicit.
Window and sample selection
Record how the entry and exit dates were chosen. A date search across many combinations needs a separate validation sample or robustness check. Show completed trade count as well as the configured lookback. Compare average with median and individual annual paths.
Trading reality
Include all contract-side commissions, estimated slippage and changes in bid-ask depth near expiry. Stop the backtest before a delivery or broker deadline. A profitable close-to-close theoretical basket may not have been executable at those exact prices.
Risk and replication
Calculate worst completed loss, worst interim adverse move and variation-margin needs. Save the symbols, dates, source and code version needed to reproduce the result. If a screenshot cannot be tied to a reproducible basket, do not use it as evidence of an edge.
Research the exact spread
Compare dated legs, historical years and current market conditions before using an average curve.
Open Futures Spread Analyzer →Frequently asked questions
Is the seasonal average enough?
No. It can hide divergent years, missing observations and path risk.
Does a positive backtest guarantee profit?
No. Historical selection and future regime changes can invalidate the apparent pattern.
Exchange references
Confirm current specifications, listed combinations and margins with the exchange and broker.
Continue the Futures Spread Trading Strategies series
Also see Calendar Spread Trading, Intercommodity Spreads and Butterfly Spreads.
Test the structure yourself
Inspect individual years, trading windows and the full basket before making a decision.
Explore the Analyzer →Educational material only, not investment advice. Illustrative examples exclude trading costs unless stated. Historical results do not guarantee future performance.