Choosing Prices for Proportional Futures Rollover Adjustments

Summary

The document asks which observations should anchor a proportional adjustment when building a continuous futures series: the expiring contract’s close and next contract’s close on rollover day, or the expiring contract’s prior-day close and next contract’s rollover-day open. It provides no conclusion or worked calculation; it frames a practical data-handling decision for futures price histories.

The choice affects which contract prices are aligned at the roll and whether the adjustment incorporates the rollover session’s price movement. The note does not compare the alternatives empirically, specify a trading or backtesting objective, or discuss intraday timing and execution. Thus it identifies an important construction question but leaves the appropriate convention unresolved.

Key ideas

• Proportional adjustment can be used to link futures contracts into a continuous price series.
• One proposed adjustment uses both contracts’ closing prices on the rollover date.
• Another uses the expiring contract’s previous close and the next contract’s rollover-date open.
• The document poses the choice but does not establish which convention is preferable.

Tags

# Future Rollover - which price to use?


# Future Rollover - which price to use?












In order to generate continuous future contract price, i adopt the proportional adjustment. Suppose rollover date is `t`.

If I have tick data of the spot month contract and next month contract, should I use spot month's close price on `t` and next month's close price on `t` to compute the adjustment factor and multiply the spot month price from day 1 to `t`?

Or should I just stick to the general approach and use the spot month's close price on `t-1` and next month's open price on `t` to compute the adjustment factor and multiply the spot month price from day `1` to `t-1`?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.