Card Removal Priority: What to Cut First at the Shops

Every draw is a competition between the cards you added and the cards you started with. Starter decks are mostly Strikes and Defends, and both fade quickly. A Strike that deals a few damage still costs energy you could have spent on something that scales. A Defend that blocks a few damage sits in your hand while you look for the card that would actually have saved the turn.

Removal also makes upgrades worth more. Upgrading a card you draw twice per fight beats upgrading a card stuck behind three dead draws. When you remove a weak card, every other card comes up slightly more often, and that effect compounds as the deck shrinks.

None of this means removing everything. A deck needs enough damage and enough block to survive each turn, and a deck with no basic cards left can still lose to a bad hand. The goal is not the smallest deck. The goal is a deck where every card you draw is one you are happy to play.

Start with the cards your deck already replaces. If your attacks come in pairs and a Strike is now the weakest one, cut a Strike. If block comes reliably from a skill or a power, cut a Defend. Keep one or two of each while the deck still depends on them for an ordinary turn; cut them when the deck no longer falls apart without them.

Then deal with permanents. Curses and status cards that stay in your deck between fights cost you a draw every shuffle. If a merchant is close, buy the curse removal before anything else, because a curse keeps charging for the whole rest of the run while a card only helps when you draw it.

Then look at your own mid-game picks. Everyone takes a card in Act 1 because the deck needed damage, and by Act 3 that same card is the weakest thing in the deck. Removing it is usually cleaner than adding another card to cover for it. The deck management guide covers the same logic from the skip side.

Ironclad decks usually cut Defends once block comes from exhausting cards or from one strong defensive card. Strength scaling also makes Strikes less relevant as attacks, so the first removals are often Defends, then spare Strikes. With exhaust strategies, be careful: cheap default cards are fuel, and the strategy can want a few of them.

Silent decks can keep Strikes longer because poison and shiv plans still need some front-loaded damage. Curses and low-value skills go first. Discard plans want a steady supply of cards to throw away, so keep enough cheap cards to feed the engine instead of trimming the deck down to nothing.

Defect decks usually keep a Strike or two while orbs are still ramping, then cut Defends once Frost provides a floor. After Frost is reliable, every remaining Defend is competing with a card that would have channeled or blocked better. The Defect orb guide explains why block cards and orb setup want the same slots.

Regent and Necrobinder decks lean on cards that produce resources. Cut cards that generate a resource your deck cannot spend any more, not the cheap cards you play every turn. The principle is the same everywhere: remove the card whose job something else already does better.

The merchant is the reliable source. A removal is one of the cheapest useful things you will ever buy and it never becomes obsolete. Compare it with the cards on offer: a removal that makes the deck consistent usually beats a card that makes one turn bigger.

Events offer removals in exchange for health, gold or a transform. Take them when the price is small and the card you are deleting is a real problem. Do not trade a lot of health for a removal right before an elite you already struggle with.

The blessings that open each act sometimes include a removal or a transform. These are excellent because they cost nothing but the choice. If your deck still carries several starters, a removal can beat a flashy bonus you will not get to use. The Ancient blessings page explains the options.

Transforms are removal with a gamble attached. They give you a random card in exchange for the one you gave up, which can be great and can also be the same tier of card again. Use them on genuinely bad cards, not on cards that are only slightly below your best.

A steady climbing deck usually wants four to six removals across a full run, spread out so the deck is never left without enough block or damage for the act it is in. That is a guideline, not a rule. A deck that adds a lot of cheap cards needs fewer; a deck that plays one big card per turn wants more.

If you have gold and nothing worth buying, a removal is almost always available and almost always fine. If you are short of health, a potion or a rest is worth more in the moment. Judge by what the next two fights need, not by what looks efficient in theory.

At the end of a run, look at the deck and find the three worst cards. Ask how many fights would have gone better if each had been removed a floor earlier. That habit is what this whole guide is for.

Do not cut your last real block card because it is small. A small Defend played on the right turn is worth more than a perfect card that stays in the draw pile. The same applies to your only source of a needed resource: your last poison applicator, your last way to generate energy.

Do not remove a card you cannot replace yet just to reach a target deck size. A small deck of weak cards loses to a large deck of strong ones. Size is not the goal; every-card-does-a-job is the goal.

Do not pay for a removal you do not need. Gold kept for a potion before the boss, or for a card that fixes a real gap, is not wasted. The merchant is not a vending machine where every purchase is correct.