The spread at the signal moment is not the spread at the fill moment

[Fairman]

The spread at the signal moment is not the spread at the fill moment

A system checks the spread at signal time and sees 0.3 pips, fine to proceed. By the time the order reaches the server and executes, the spread has jumped to 0.9. The entry happens at a worse price than assumed, and the trade starts underwater.

This is especially common right when a signal triggers, since the same price movement that caused the signal often triggers liquidity providers to widen their quotes. The signal and the spread spike are correlated, which makes the problem systematic and not random.

To measure the effect, log both the spread at signal time and the spread at fill time. The average difference reveals how much extra cost is hidden. Many traders find it adds a significant fraction of a pip.

Mitigation options include using limit orders with a maximum acceptable price, rechecking the spread immediately before sending, or building the expected spread jump into the target.

The key point is that cost estimates based on quiet moments understate the true cost of acting on fast signals.

How large was the gap between signal spread and fill spread in your own logs?