One weighted model, nineteen commander configurations, ranked. This page consolidates the scoring built across your ChatGPT chats and audits whether every chat is working from the same nineteen. It does not replace the live marvel-commander-rankings page.
Each commander carries five sub-scores on a 0–10 scale. The overall is their weighted sum. The weights total exactly 1.00, so the overall stays on the same 0–10 scale.
overall = 0.28·power + 0.26·ceiling + 0.24·fun + 0.12·ease + 0.10·politics → e.g. Captain America 7.9/8.2/8.1/8.2/5.5 = 7.82
COMMANDER_RANKINGS with byte-identical slugs, order, and every overallRaw. The gaps are the live page and the big review chat, both flagged below.
The live page keeps the older {role, nm, cost} deck model with 45 rows (7 face + 3 second + 35 alt). The canonical 19 is a separate, newer COMMANDER_RANKINGS payload not yet wired in. Reconciling the page is a deliberate next step, held pending your go–ahead (you asked not to overwrite it).
Click any row to reveal its one-line rationale. Sorted by overall, descending. Confidence reflects EDHREC sample size — the tail (She-Hulk, Ms. Marvel, T'Chaka) rests on single-digit public decks and is marked provisional.
| # | Commander | Deck | Overall | Pow | Ceil | Fun | Ease | Pol | EDHREC | Conf. |
|---|
Separate from the commander scores above: these are the quality grades ChatGPT gave each live page during the review session. Kept distinct so the two 0–10 scales are never conflated.
Every number here traces to a captured chat, saved byte-for-byte. Nothing is invented.
~/Projects/_outputs/marvel-critique/ — Marvel Commander Rankings Review (26), Commander Rankings Review (94), Commander 2026 Card Upgrades (15), Team Leader Ranking Confirmation (17), Commander Ranking Analysis (9)marvel-19-commander-scoring.json (verified, weights sum 1.0)