Why quote timestamps matter more than screenshot order

Quotes obtained at different times can differ because the market changed. Record when each quote was observed and, where available, its underlying block or state. Screenshot order is not a controlled comparison of execution quality.

A timing confound

At 12:00:00, hypothetical Route A quotes 1,000 output tokens. At 12:00:15, Route B quotes 1,004. If the shared market improved by roughly 0.5% during that interval, B's larger output does not prove that its route was better. A contemporaneous A quote might have improved too.

The reverse problem also occurs: a route can appear worse simply because it was checked after an adverse market move. Do not attribute every between-screen change to the aggregator.

Use a defensible observation protocol

Request matched amounts as close together as practical. Preserve each quote time, the expiration and any state identifier exposed. If quotes take different amounts of time to return, distinguish request time from response time.

Pool execution prices can change as reserves change. Matching only the calendar date or minute is therefore insufficient for some comparisons, especially when the apparent advantage is small.

What a timing window permits you to claim

You can report the best observed quote in a specified window. That is weaker than claiming the best simultaneous executable quote. If one quote has already expired when the other arrives, they may never have been available together.

For a retrospective evaluation, keep the actual observations and state the timing gap. An external reference series can help identify broad market movement, but it cannot reconstruct every unavailable route quote. Repeated observations improve coverage only when the collection method is consistent; collecting many unmatched screenshots does not remove timing bias by itself.