say what the Solana leader's number is, and count the providers right - #2815
Merged
Merged
Conversation
Florent read "serialized 0.001 s" off the Solana column and asked whether that was real. It is the 1 ms floor: the feed that wins a race has a lag of zero by definition, published as 1 ms so the row renders at all, because the site drops rows whose headline is exactly zero. The methodology states the floor. It does not state the consequence, and the consequence is the part a reader acts on: the leader's figure is a constant, so the column is not a latency for whoever is winning, and the gap to the next feed is not a speed ratio. Measured over 30 minutes to put a number on it: provider p50 p90 on the 1 ms floor serialized 0.001 0.001 94.4% codex 0.082 0.102 15.0% mobula 0.110 0.325 32.2% okx 0.236 0.292 0.0% birdeye 0.772 1.007 0.0% This is the same mechanism behind an earlier false alarm in this work, when a single instantaneous read showed birdeye at 0.001 on Solana and I reported it as selection bias. It was the floor then too. And the title said "across 5 data providers", rendered three times on the aggregate page, while six are live. A factual correction rather than a retitle: the name and keyword phrasing are untouched, only the count. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Flotapponnier
added a commit
that referenced
this pull request
Oct 4, 2026
Spec only, from #2815: the bullet saying the leading feed's Solana figure is the 1 ms floor rather than a measured latency, and the provider count in the title corrected from 5 to 6. The worker clones dev at BUILD time, so the methodology reaches the page from here. pnpm validate: 235 specs valid. Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Florent read
serialized 0.001 soff the Solana column and asked whether that was real.It is the 1 ms floor. The feed that wins a race has a lag of zero by definition, published as 1 ms so the row renders at all, because the site drops rows whose headline is exactly zero (
raceLagSeconds).The methodology states the floor. It does not state the consequence, and the consequence is the part a reader acts on: the leader's figure is a constant, so the column is not a latency for whoever is winning, and the gap to the next feed is not a speed ratio.
Measured over 30 minutes:
Same mechanism as an earlier false alarm in this work: a single instantaneous read showed birdeye at 0.001 on Solana and I reported it as selection bias. It was the floor then too.
The provider count
titlesaid "across 5 data providers", rendered three times on the aggregate page, while six are live. A factual correction rather than a retitle: the name and the keyword phrasing are untouched, only the count, so no ranking history is reset by a reworded title.pnpm validatereports 235 specs valid.🤖 Generated with Claude Code