Bant

A slot is one card in the average list. For each card, we count the lists that play at least one copy, then at least two, then at least three. We add those shares. Add every card and the boards total 60 and 15.

The settled bound is chosen at 80% of published lists. Settled slots are grey. Other slots are violet. Cards in fewer than 10% of lists still count in the argued-over total, but their rows are not shown.

The band bounds are chosen. Locked starts at 60 settled slots out of 75. Consolidated starts at 50. Debated starts at 40. Open is below 40. A locked deck is not a better deck.

Share is share of published decklists, not of the field. Each list counts once, paper or MTGO. No platform and no event is weighted. An Arena event we add by hand counts here too.

The two platforms publish different lists. A paper event on Melee publishes almost every player's list. An MTGO Challenge publishes 32 lists, whatever its size, and those lists lean to the top of the standings. So paper and MTGO together mix almost the whole paper field with MTGO's top finishers. One big paper event can be a large part of a window. Pick MTGO or Paper to see one platform alone.

Leagues publish 5–0 decklists and no match data. Deck counts and shares count them, because a registered list is a registered list. Win rates cannot, because a League has no matches. So counts and shares cover all published decklists, and win rates cover events with match data only.

Preliminaries contribute nothing. They publish match records but no decklists, so their matches have no archetype to attach to.

Win rates count only matches where both players published a decklist. With both sides counted, every win has a matching loss, so the pool always balances at exactly 50% wins.

On paper, almost every match has both lists, so a paper win rate measures a deck against the whole field. On MTGO, the published lists are the top finishers, so an MTGO win rate measures a deck among the top finishers. Paper and MTGO together mix the two.