feat(cast/explain): candidateScoreNonWinnerQMGap (76767676)

[chitcommit]

Summary

• Adds candidateScoreNonWinnerQMGap to buildCastExplanation(): absolute gap between non-winner quadratic mean and arithmetic mean (QM - AM)
• Present whenever >= 3 candidates — no zero restriction
• Always >= 0 by Power Mean inequality; zero when all non-winners tied
• Variance identity: QMGap * (QM + AM) === stdDev^2 (from QM^2 - AM^2 = stdDev^2)
• For n=3: sqrt((r^2+t^2)/2) - (r+t)/2
• Additive-scale complement to candidateScoreNonWinnerQMToAMRatio

Test plan

• 76767676-1: present when >= 3 candidates
• 76767676-2: always finite and >= 0
• 76767676-3: identity — gap + AM === QM
• 76767676-4: for n=3 equals sqrt((r^2+t^2)/2) - (r+t)/2
• 76767676-5: absent on cast:no_match
• 76767676-6: absent when fewer than 3 candidates
• 76767676-7: variance identity — gap*(QM+AM) === stdDev^2
• 76767676-8: tool description documents candidateScoreNonWinnerQMGap

https://claude.ai/code/session_01VKBFeSfC62NKttk4spVHsB

---

Generated by Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[coderabbitai]

Caution

Review failed

An error occurred during the review process. Please try again later.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch auto/76767676-cast-explain-non-winner-qm-am-gap

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}