A sideboard is a matchup plan

People build sideboards by collecting answers. Rest in Peace, Wear // Tear, the extra discard spell, a creature that is good against a strategy they hate. Then they sit down, look at fifteen cards, and guess.

A sideboard slot needs two sentences: what it comes in against, and what it replaces. If you cannot name the cut, the card is not a sideboard card. It is a wish.

Start from the matchup, not from the binder. Against a graveyard deck you need a plan for their engine and a plan for the game if the engine is gone. One hate card that you cannot protect is a prayer. Two pieces, or one piece plus a way to draw it, is a plan. Against a faster deck you need cheap interaction and a lower curve — which means the cuts are your slowest cards, even if they are your most powerful in a vacuum.

Number of copies matters. One-ofs are for cards you never want two of, or for a toolbox you can search. If you need to see the effect by game two on the play, one copy is usually not enough.

After a tournament, the sideboard is the first thing to autopsy. Which cards never came in? Which matchups were you still “figuring out” after boarding? Those fifteen slots should be as deliberate as the sixty. The sixty is your default plan. The fifteen is how you refuse to play that plan when it is wrong.