fix(ftip): distinguish archive maximizers and correct reliability arithmetic [AGENT] - #179
Merged
Conversation
…thmetic [AGENT] Define finite real whole-chain means on a nonempty task set and archive, and separate attainment from deployment eligibility. State selector membership and prove the finite approximation bound. Replace the disjoint-strengths claim with exact whole-chain and per-task oracle calculations, and correct aggregate agreement to 99/101 with zero safety-subgroup agreement.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The archive example conflated the maximum of whole-chain mean scores with a per-task oracle. Keep the finite archive maximum, state its real-valued domain and attained maximizer, and separate eligibility under a named deployment configuration. The two-chain example now computes 1/2 for either fixed chain and 1 for task-dependent oracle selection; any implemented router remains a separately evaluated policy with its costs.
The selector bound now states membership in the same archive and a nonnegative approximation tolerance, with its proof. The grader example gives 99/101 overall agreement and zero agreement within the safety subgroup. Stable addresses, the wrapper/archive/reliability layers and neighboring domain conventions are preserved.
Validation: exact mathematical controls, all 22 existing prose tests and lint, full local build, rendered-output/prose checks, nine browser routes, and a visually inspected 237-page focused PDF passed. All 773 source files, 773 HTML payloads and the PDF are archived and verified against the candidate. A fresh independent adversarial review approved exact head
01df43cc4d0d1a7cd4dab29602c663b6f8094246with no findings.