Read deck statistics without fooling yourself

A percentage needs context. Look for the patch, ranks, exact lists, and game count. Without these, the number may answer a different question from yours.

What card stats show

• Include rate: How often a card appears in the chosen decks. This can show popularity as well as power.
• Kept percentage: How often players keep it in the mulligan. Players can make poor keeps too.
• Drawn win rate: Results in games where it was drawn. Game length, matchups, and deck lists can all affect this.
• Played win rate: Results when it was played. A finisher may score high because people play it just before winning.

Adding another finisher does not always improve a deck. You may need the early defense you would cut for it.

Mulligan win rate needs the source's exact definition. Some count cards offered at the start, even when replaced. Others use different filters. Check before comparing sites.

Going first and second are different groups. Hand size, the first play, and The Coin all matter.

Matchups and rankings

One deck name may group very different lists. A small sample may include many skilled players or an unusual mix of opponents.

A high-Legend result proves that one player reached that result with the list. It does not prove a win rate for all players.

A patch is a reason to check the data again. Mixing games from before and after a change can hide its effect. Keep older reports as history, not live rankings.

Chance and choices

Expected value is an average that gives more weight to more likely results. Variance describes how widely results can differ.

When ahead, a steady plan may protect your lead. When far behind, a risky plan may keep one winning chance open. Good luck does not prove a good choice. Review what you knew before the result.