Skip to content

feat(napkin-math): sharper severity wording in Suggested next actions + MARGINAL bucketing#722

Closed
neoneye wants to merge 1 commit into
mainfrom
feat/napkin-math-suggested-actions-and-marginal-wording
Closed

feat(napkin-math): sharper severity wording in Suggested next actions + MARGINAL bucketing#722
neoneye wants to merge 1 commit into
mainfrom
feat/napkin-math-suggested-actions-and-marginal-wording

Conversation

@neoneye
Copy link
Copy Markdown
Member

@neoneye neoneye commented May 17, 2026

Summary

ChatGPT review of the v44 casino_royale assessment flagged two wording issues — both in summarize_assessment.py, both pure rendering, no schema change.

1. Suggested next actions item #1 understated DOOM

Old:

"1 gate(s) currently fail at the 50% pass-rate bar."

That phrasing reads identically whether the worst gate is FRAGILE at 49% or DOOM at 0%. ChatGPT: "the worst gate has 0.0% pass rate, which is not just 'below 50%'; it is a structural failure under current bounds."

New (mixed example):

"2 declared gates in the DOOM band; 3 in the FRAGILE band. Worst: sponsor_profitability_window_margin_days at 0.0% pass rate under current bounds."

New (single-DOOM example, matches casino_royale v44):

"1 declared gate in the DOOM band. Worst: sponsor_profitability_window_margin_days at 0.0% pass rate under current bounds."

The summary distinguishes DOOM count from FRAGILE count and names the worst gate by id and pass rate.

2. MARGINAL "coin-flip" wording over-fired at 79.8%

Old Decision implications for MARGINAL was a single template:

"passes 79.8% of runs — close enough to coin-flip that downstream commitments should not assume it holds."

ChatGPT: "At 79.8%, it barely misses ROBUST. Calling it 'close enough to coin-flip' is too harsh. It is not coin-flip; it is near the ROBUST threshold."

The MARGINAL band is 50–80%. A 51% pass is genuinely coin-flip; a 79% pass is one slip from ROBUST. Bucketed at 70%:

  • ≥ 70%: "just below the ROBUST band. The gate passes in most runs, but downstream commitments should not treat it as secure."
  • < 70%: keeps the "close to coin-flip" framing.

Test plan

  • Smoke 9/9, unit 50/50 green.
  • v44 casino_royale regenerated locally: item fix: Add UTF-8 encoding for markdown file operations #1 now reads 1 declared gate in the DOOM band. Worst: sponsor_profitability_window_margin_days at 0.0% pass rate under current bounds. and the 79.8% AML-adjusted NOI gate's Decision implications row now reads just below the ROBUST band. The gate passes in most runs, but downstream commitments should not treat it as secure.

🤖 Generated with Claude Code

…+ MARGINAL bucketing

ChatGPT v44 review: two wording fixes.

Suggested next actions item #1 previously said 'N gate(s) currently fail at the 50% pass-rate bar' regardless of whether the worst pass rate was 0% or 49%. That phrasing understated DOOM failures: '1 gate fails at the 50% bar' reads identically whether the gate is FRAGILE-48% or DOOM-0%. Now distinguishes DOOM vs FRAGILE counts and names the worst gate by id + pass rate: '1 declared gate in the DOOM band. Worst: sponsor_profitability_window_margin_days at 0.0% pass rate under current bounds.' (or '2 in the DOOM band; 3 in the FRAGILE band. Worst: ... at X.X% pass rate.' for mixed cases).

Decision implications MARGINAL wording was 'close enough to coin-flip' across the full 50-80% band. At 79.8% that reads as a misdiagnosis — the gate is one slip from ROBUST, not coin-flip. Bucketed at 70%: at-or-above 70% uses 'just below the ROBUST band. The gate passes in most runs, but downstream commitments should not treat it as secure.'; below 70% keeps the 'close to coin-flip' framing.

No schema bump (manifest unchanged; pure rendering). Smoke 9/9, unit 50/50.
@neoneye
Copy link
Copy Markdown
Member Author

neoneye commented May 17, 2026

Superseded by #723, which bundles these two wording fixes with the v45 saturated-gate-exclusion explainer.

@neoneye neoneye closed this May 17, 2026
@neoneye neoneye deleted the feat/napkin-math-suggested-actions-and-marginal-wording branch May 17, 2026 02:40
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant