The price moved first. Your fill may arrive last.

Zoom into two price series and you may spot a leader and a follower. It is tempting to assume you can buy the follower before it catches up.

The chart does not show the full clock.

Measure every delay

Your feed has a publishing delay. Your connection adds another. The strategy needs time to calculate, sign, submit, and receive a response. The market may move during all of it.

Align the source timestamps before comparing series. Two feeds can appear to lead and lag simply because one labels the start of an interval and the other labels the end.

Replay at obtainable prices

A backtest that buys the displayed midpoint at the exact signal time is usually making a generous assumption. Use the ask available after a realistic delay, and limit the size to available depth.

Test several delay settings. If a strategy becomes unprofitable after a tiny increase, its apparent edge may be too fragile for the intended setup.

Include periods with feed outages and stale observations. Those are part of the operating environment, not inconvenient rows to drop.

Speed can matter. So can declining an order when the quote is already gone.

The best lead-lag study ends with a distribution of actual fills and missed opportunities. A pair of lines on a chart is the hypothesis.

Further reading: Prices and order books · Resolution rules · Current fees.

Explore the markets: Open Polymarket (referral link).