Futures rollover: why a continuous chart is not your traded contract.

A future is always a specific contract

NQ, ES and CL are product families. What actually trades is a dated contract such as NQZ6. Each contract has its own order book, volume and price. A root symbol without a month is not a position you can execute.

That distinction matters in a review. Entry, exit and candles must refer to the same contract. Plotting a September fill on December candles can create an apparent mismatch even when both data sources are correct.

Rollover and expiration are different events

CME describes rollover as moving exposure from an expiring contract to a later month. Traders commonly watch volume in both contracts to decide when to switch. Expiration is the end of the individual contract. A position that is neither offset nor rolled proceeds toward cash settlement or physical delivery, depending on the product.

For U.S. equity index futures, CME lists the customary roll date as the Monday before the third Friday of the expiration month. After that date, the second-nearest quarterly contract is commonly treated as the lead month. This is a liquidity convention, not a forced platform switch.

Why a continuous chart may smooth or create a gap

A continuous series joins several contracts. Because two maturities do not have to trade at the same price on the switch date, the data provider needs a rule: join them unchanged, back-adjust history or apply a ratio. Different methods can produce different historical levels.

Continuous data is useful for broad market structure. To determine whether your entry, stop or target actually traded, use the exact contract and unadjusted intraday candles.

Simple rule

Continuous for context. The traded contract for execution review, fill reconciliation and bar-level testing.

Avoid false backtest results around the roll

Store the full provider symbol and contract month for every trade. Load candles for that exact symbol across the entry-to-exit window before testing whether an alternative target or stop would have traded.

Longer backtests need a roll rule defined in advance. A strategy tested on back-adjusted data may use levels that were different in the live front month. That does not automatically invalidate the test, but the methodology must be disclosed.

• Store provider symbol and normalized product separately.
• Never infer the roll from the calendar month alone.
• Check volume migration and the official product calendar.
• Use the actually traded contract for execution reviews.
• Document the adjustment method for continuous data.

A five-point check before each quarterly switch

Confirm the current lead month, compare volume with the next contract, check the last trading day, update watchlists and verify the symbol in broker, chart and journal. Energy, metal and agricultural futures use different expiration and delivery rules from equity indexes.

The central lesson is deliberately plain: the symbol is not metadata. It decides whether the chart, execution and backtest describe the same market observation.