docs(README): replace 'honest open question' with measured 80% real-app precision#267
Merged
Merged
Conversation
…recision Per the spoonworks v0.52 benchmark, real-app precision is now 4/5 = 80 % (measured 2026-05-14, full triage in docs/benchmarks/BENCHMARK_SPOONWORKS.md). Baseline was 6/77 = 7.8 %; the three-step trajectory is now in the README so the empirical numbers tell the story rather than aspirational framing. Also reframes 127/127 self-test number with the explicit caveat that it's synthetic — real-app number is load-bearing. Dropped the v0.50 'open question' paragraph (it's no longer open).
|
✅ BugHunter Calibration | | 2026-05-14 Overall: tp=0 fp=0 fn=0 precision=1 recall=1 f1=0
|
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Real-app precision on spoonworks (measured 2026-05-14 with v0.52 from PR #266) is 4/5 = 80%, a 10× uplift from baseline 7.8%. README now leads with this number instead of the synthetic 127/127 framing.
Why
The prior README said "the current FP rate on real apps is the honest open question — not the kind-recall number." That sentence is no longer true: the spoonworks benchmark (docs/benchmarks/BENCHMARK_SPOONWORKS.md) measured precision empirically, and the v0.52 fixes (PRs #265, #266) brought it to 80%.
Trajectory shown in README
The single remaining unexplained cluster (`missing_state_change` on a "Remove row" button) appears to be a detector over-fire — manual code review of `components/admin/inventory/RecipeEditor.tsx` confirms the handler removes the row from state correctly. Tracked as a follow-up.
Also
🤖 Generated with Claude Code