Which cards earn their slot? Reading card utilization

Every deck has a card the pilot defends in theory and never actually uses. Finding it by feel takes months; finding it in data takes a decent sample of logged games. If you track what happens to each card across your games, four rates tell the story.

The four rates

• Drawn — games where you drew the card at least once.
• Played — games where you actually used it.
• Discarded — games where it hit the discard pile (as a cost, from your hand, however it got there).
• Dead in hand — games where you drew it and it was still in your hand, unused, when the game ended. This is the loudest signal on the table.

Count games, not copies: drawing two Ultra Ball in one game is one game where Ultra Ball was drawn. Per-game rates are what answer the deck-building question — "when this card shows up, does it do anything?"

Read the rates through the card's role

The same numbers mean different things for different cards, so judge each card against its job:

• Core-job cards — attackers, your draw and search engine, the card the deck is built around — exist to be used. One of these sitting dead in a quarter of your games, or drawn often but rarely played, is a genuine cut candidate or a sign the list can't support it.
• Situational cards — Stadiums, tech Items, counter-cards — idle by design; that's part of their job. Judge them gently. Only an egregious dead-in-hand rate, or a matchup where they were supposed to shine and didn't, earns a mention.

One role no database can assign for you: the key card your deck is actually built around. Tools can auto-detect draw and search text; only you know what the deck is for.

Two honesty rules before you cut anything

First: "no data" is not a cut signal. A card with no logged games has earned no verdict, good or bad. Second: per-card rates need more games than deck-level rates — on the order of 30 logged games before a card's numbers deserve a list-surgery decision. The dead-in-hand rate of a one-of after eight games is an anecdote, not evidence.

Close the loop with theory

Utilization is the practice half of a loop. Draw math told you what each copy count should deliver; the theory band told you what setup rate to expect. Utilization tells you what each slot actually bought. A card that's mathematically likely to show up and empirically does nothing when it arrives is the cleanest cut in deck building.