Back to Tips
Strategy GuideVerified March 17, 2026

Deck Thinning & Card Removal Strategy

Complete guide to deck thinning in Slay the Spire 2 — ideal deck sizes by character, card removal priority order, removal cost scaling, when NOT to thin, and removal vs transform trade-offs.

Removing a Strike from your deck does more for your win rate than adding most common cards. Every card you cut increases the probability of drawing the cards that actually win fights. A 20-card deck draws each card 25% of the time over four turns. A 30-card deck draws each card 17% of the time. That gap compounds across every fight in a run.

Deck thinning is not about reaching some magic number. It is about reaching the point where every card in your draw pile contributes to your game plan every time you see it. This guide covers the concrete numbers — removal costs, ideal deck sizes by character, priority order, and the specific situations where thinning is wrong.

Verification note

Removal costs and scaling confirmed across ScreenHype, GAMES.GG, and Mobalytics beginner guides. Deck size ranges cross-referenced with SlashSkill archetype guide, Mobalytics character guides, and community consensus. Insatiable boss interaction verified via GamerBlurb, Phrasemaker, and NeonLightsMedia boss guides.

Fast takeaway

This guide is built around one practical question, so you can use it during a run instead of digging through a broad overview.

If the answer depends on a mechanic, a character system, or a recent patch, the related links show you what to open next.

Use this when you want a direct answer instead of a broad overview.

Follow the related links if this decision depends on a mechanic, character system, or co-op rule.

Check the update pages whenever balance changes might shift the recommendation.

Why consistency beats raw power

A deck with 15 strong cards and 10 basic Strikes and Defends is weaker than a deck with just those 15 strong cards. The Strikes and Defends do not add power — they dilute it. Every draw spent on a basic card is a draw not spent on your scaling, your damage combo, or your key block card.

The math is direct. In a 25-card deck, your 5-card opening hand contains any specific card 20% of the time. In a 15-card deck, that same card appears 33% of the time. For cards like Limit Break, Catalyst, or No Escape — cards that define whether you win a fight — that difference between 20% and 33% is the difference between a clean kill and taking 40 damage while waiting to draw your win condition.

Ideal deck sizes by character

There is no universal ideal deck size. Each character's mechanics create different thresholds where deck size transitions from 'lean and consistent' to 'bloated and unreliable.'

  • Ironclad (25–30 cards): Ironclad decks run larger because Exhaust effects like Burning Pact and True Grit act as in-combat deck thinning. A 28-card Ironclad deck with Dark Embrace and Feel No Pain plays like a 15-card deck after the first two turns of exhausting. Ironclad also benefits from having a wider spread of block and damage options because Strength scaling makes even mediocre attacks hit hard late in a fight.
  • Silent (20–25 cards): Silent decks need to draw Catalyst or key Sly cards consistently. Poison builds want to see Catalyst and Noxious Fumes every fight. Sly/discard builds need their discard triggers every turn. Both archetypes reward tight decks in the 20–25 range. Shiv builds can tolerate slightly larger decks because Shivs generate their own card draw through After Image and Kunai triggers.
  • Necrobinder (15–20 cards): Necrobinder benefits the most from aggressive thinning. Souls are 0-cost Exhaust cards that draw 2, meaning a thin deck cycles faster through its draw pile. Forbidden Grimoire provides free post-combat removal every fight. A 15-card Necrobinder deck with Haunt and a Soul engine can cycle its entire deck every turn, creating pseudo-infinite damage. This is the character where removing every possible basic card pays the highest dividend.
  • Defect (20–25 cards): Defect decks revolve around finding specific combo pieces — Claw + All for One + Scrape, or Electrodynamics + Focus generators. Consistency matters enormously because Defect's damage ceiling depends on assembling the right hand, not on raw card power. Keep decks tight to hit Claw loops or Orb setup turns reliably.
  • Regent (varies, 20–30 cards): Regent is the exception. Forge builds often run larger decks because Forged cards are individually powerful enough to carry turns even without specific draw order. Stars builds want tighter decks (20–25) to hit Big Bang and Alignment consistently. If your Regent deck mixes Stars and Forge, lean toward 25 cards and prioritize draw cards to offset the size.

Card removal priority order

Not all cards are equally worth removing. The priority order reflects which cards provide the least value per draw as your deck evolves through a run.

  • Priority 1 — Strikes: Remove Strikes first. A basic Strike deals 6 damage (9 upgraded). By mid-Act 1, almost every card you pick up deals more damage, has additional effects, or scales with your build's mechanic. Strikes are pure dead weight. Removing one Strike has a larger positive effect on your average hand quality than adding one uncommon card.
  • Priority 2 — Defends: Remove Defends after Strikes are gone. A basic Defend gives 5 Block (8 upgraded). Class-specific block cards like Shrug It Off (Ironclad, Block + draw), Backflip (Silent, Block + draw), and Glacier (Defect, Block + Frost Orb) do the same job but with extra value. Defends stay slightly longer than Strikes because 5 Block in a bad hand is less dangerous than 6 damage in a bad hand.
  • Priority 3 — Early picks that no longer fit: Cards that solved Act 1 problems (single-target damage for hallway fights, basic AoE) but do not contribute to your Act 2–3 scaling plan. If you picked up a Cleave to handle Act 1 multi-enemy fights but your deck now revolves around Strength + Heavy Blade, that Cleave is a dead draw.
  • Priority 4 — Curses and Afflictions: Curses always deserve removal when offered. Afflictions attached to existing cards are harder to address — they cannot be removed independently; the afflicted card itself must be removed or transformed. Prioritize removing afflicted cards that were already marginal.

Removal cost scaling and economy relics

Shop card removal starts at 75 gold and increases by 25 gold for each removal purchased during the run. The second removal costs 100, the third costs 125, and the fourth costs 150. This scaling makes early removals dramatically cheaper than late removals.

  • First removal: 75 gold. Almost always worth buying. Removing a Strike in Act 1 improves every fight for the rest of the run.
  • Second removal: 100 gold. Still efficient. Two removals in Act 1 (removing both Strikes from a shop visit) costs 175 gold total — less than most rare relics and more impactful than most uncommon relics.
  • Third removal: 125 gold. Evaluate against other shop options. At 125 gold, a removal competes with buying a strong uncommon card or a useful potion. Remove if the worst card in your deck is a clear liability; buy a card if nothing in your deck is actively bad.
  • Fourth removal: 150 gold. Only remove if the target card is a curse, an afflicted card, or you are specifically building toward a sub-20 card deck.
  • Smiling Mask: Sets removal cost to a flat 50 gold regardless of how many removals you have purchased. The scaling reset makes this relic worth 25 gold on the first removal, 50 gold on the second, 75 gold on the third, and more on every subsequent one. If you plan to remove 4+ cards (Necrobinder and Silent builds often do), Smiling Mask is one of the strongest economy relics in the game.
  • Membership Card: Reduces all shop prices by 50%. With Membership Card, removal costs drop to 38–75 gold across the scaling range. However, Smiling Mask overrides Membership Card for removal pricing — if you have both, removal stays at 50 gold flat, not 25.

Removal channels beyond the shop

The shop is the most reliable removal method, but events and card effects also thin your deck. Recognizing these opportunities lets you preserve gold for other purchases.

  • Events: Several Unknown room events offer card removal as a reward. These free removals are always valuable. When pathing on the map, routing through Unknown rooms increases your chance of hitting a removal event, especially in Act 1 where the event pool contains multiple removal options.
  • Exhaust effects (in-combat thinning): Burning Pact (Exhaust 1 card, draw 2), True Grit (gain Block, Exhaust a random card), and Second Wind (Exhaust all non-Attack cards, gain Block for each) thin your deck during combat. Exhausted cards do not return to your draw pile for the rest of the fight. Ironclad and Necrobinder decks can use Exhaust as a pseudo-removal engine that makes a 28-card deck play like a 15-card deck.
  • Forbidden Grimoire (Necrobinder): This Power card lets you permanently remove one card from your deck after every combat for the rest of the run. Obtained early, Forbidden Grimoire can remove 10+ cards across a full run — more than any number of shop visits. It is the single strongest deck-thinning card in the game.
  • Peace Pipe (relic): Adds a Toke option at Rest Sites that lets you remove a card instead of resting or upgrading. Trading an upgrade for a removal is worth it when your worst card actively hurts you (curses, afflicted cards). It is less efficient than upgrading when your worst card is merely mediocre.

When NOT to thin your deck

Aggressive deck thinning is correct in most runs, but there are specific scenarios where keeping a larger deck is the right call.

  • Against the Insatiable (Act 2 boss): The Insatiable adds 6 Frantic Escape cards to your deck during the fight and places a Sandpit timer that kills you if it reaches 1. You need to draw and play Frantic Escape cards to survive. In a 15-card deck, those 6 escape cards make up 29% of your total deck — you draw them reliably. In a 40-card deck, they make up 13% — you might not see one before the timer kills you. However, the danger with extremely thin decks (under 12 cards) is that status cards added during the fight can overwhelm your hand. The sweet spot against Insatiable is 18–22 cards: thin enough to draw Frantic Escape consistently, thick enough to absorb status card pollution.
  • Curse-heavy runs: If multiple bosses and events have loaded your deck with curses and you lack Exhaust removal, a larger deck dilutes the curse ratio. A single curse in a 15-card deck appears in 33% of opening hands. That same curse in a 30-card deck appears in 17% of opening hands. When curse removal is not available, adding good cards can be better than removing mediocre ones.
  • Status-heavy Act 3 fights: Several Act 3 enemies add Burns, Wounds, Dazed, or Void cards to your draw pile. A 12-card deck that receives 5 status cards becomes a 17-card deck where 29% of draws are useless. A 22-card deck receiving the same 5 statuses becomes a 27-card deck where only 19% of draws are bad. If your deck lacks Exhaust or status-handling mechanics, maintaining a slightly larger deck provides natural resilience.
  • Infinite combo decks (special case): Infinite combos require a specific set of cards (often 5–8) and nothing else. Once you are committed to an infinite build, you remove everything except the loop components. But if the infinite is not yet assembled, premature thinning can leave you without enough cards to survive normal fights while you search for the remaining pieces.

Removal vs Transform

Several events offer a choice between removing a card and transforming it. Transforms replace a chosen card with a random card of the same rarity from your character's pool. The decision depends on the card you are targeting and where you are in the run.

  • Transform a Strike: Almost always better than removing it. The worst possible transform outcome is another mediocre common — which is still better than a Strike. The best outcome is a build-defining rare. Transforms give you a random card for free; removals give you nothing. In Act 1, always prefer Transform over Remove when the target is a Strike.
  • Transform a Defend: Slightly riskier. Some random commons are worse than a Defend in certain builds (a random Attack in a block-focused deck adds nothing). Still, the expected value of a transform exceeds the expected value of a basic Defend in most situations.
  • Transform a curse or afflicted card: Never transform a curse — you risk getting another bad card. Remove curses outright. For afflicted cards, removing is safer because the Affliction is destroyed along with the card.
  • Late-run transforms: By Act 3, your deck is more focused and a random card is more likely to be off-plan. Late-run removals are safer because they do not risk adding a card that dilutes your synergies.

FAQ

Should I always remove at the first shop?

Almost always, yes. The first removal costs 75 gold and permanently improves every future fight. The only exceptions are when the shop has a rare card or relic that perfectly completes your build, or when you entered the shop with less than 100 gold and need to save for a known upcoming shop with better offerings.

Is 10 cards too small for a deck?

For most builds, yes. A 10-card deck draws its entire contents in two turns, which means you see everything twice before reshuffling. That sounds good until an enemy adds 3 status cards — now 23% of your draws are useless. Decks below 12 cards are viable only for infinite combo builds that end fights before status cards matter.

Does Forbidden Grimoire stack with shop removal?

Yes. Forbidden Grimoire removes one card after every combat, and shop removal is a separate system with its own cost scaling. Using both lets Necrobinder strip a deck down to 10-12 cards by mid-Act 2 — faster than any other character can achieve.

Should I remove cards in co-op?

Yes. Deck thinning is equally important in co-op. Each player manages their own deck independently. Coordinate shop visits so the player who needs removal most visits the shop first, especially if only one shop node is accessible on the map.