How these numbers work
The Command Center's rule is correctness over volume: every number traces to stored observations, sources are never blended, and anything without enough sample says so instead of pretending.
Data sources stay separate
EDH popularity and card adoption read the public Archidekt deck population (recently-updated decks — an activity sample, not a census). Competitive results come from tournament feeds (TopDeck.gg pairings, MTGTop8 lists). MTGMeta users' own decks power personalization only; they are never blended into the public trends, so growth in our user base can't masquerade as a metagame shift.
cEDH means verified competitive
A generic "Commander" tournament could be a kitchen-table league or a cEDH open. Events enter cEDH statistics only with positive evidence — an explicit cEDH label or a competitive-series match; everything else is excluded until classified, not assumed in. That keeps the win rates honest at the cost of a smaller sample, which is the right trade.
Momentum
Momentum measures how fast interest is changing relative to a commander's own recent baseline, normalized by total activity (so a scraper outage moves nobody) and shrunk toward neutral for small samples (so three hobbyist decks can't outrank a 2,000-deck commander's steady climb). Brand-new commanders grow from zero by definition — they get their own "Hot New Commanders" list instead of unfairly topping the risers.
Adjusted win rates
Multiplayer pods don't have a 50% baseline: a four-player pod expects 25% wins. Every entry's expectation is computed from its actual pod size (and each top cut from its event's actual cut structure — a Top 8 of 32 expects far more cuts than a Top 16 of 200). Displayed "adjusted" rates shrink small samples toward that structural baseline; raw rates stay visible alongside their sample size.
Core, Signature and Spicy cards
Card categories compare a commander's current lists against decks of other commanders whose color identity could legally play the same cards. Core ≈ in nearly every list; Signature ≈ far more common here than in comparable decks; Spicy ≈ rare everywhere else but clearly at home here. All three require a minimum tracked-deck sample.
"Replacing" vs "frequently add"
Swap patterns are mined from deck-update transactions: among updates that removed card A, the share that added card B. One 15-card overhaul counts once, not 225 times. The word "replacing" is reserved for very strong patterns; weaker ones say "frequently add".
Known limits
Deck counts are tracked-population counts, not global censuses. Historical duplicate-merging before August 2026 means some older commander decklists were merged when they shared 99 cards; commander-aware hashing fixed this going forward. Where a number would be misleading, we show "not enough data yet" instead.
Current tuning
Momentum compares the last 7 days against the 28 before them (shrinkage K = 5). Win rates need 5 entries to display at all (prior strength S = 20). Card categories need 20 tracked decks. "Replacing" language needs 80% association across 100+ deck updates; weaker patterns say "frequently add".
