SteamWorld Quest looks like a game where the best cards are the expensive ones. Every card game trains you to think that.
The players who report clearing the late chapters without difficulty are running almost nothing that costs more than 1 Steam.
That is the whole thesis, and the maths behind it is not subtle.
One card beats the entire expensive half of the pool
Fire Pillar, upgraded twice, deals roughly 270% magic damage for 1 Steam.
Compare that to any 2- or 3-cost skill. You are spending two or three times the resource for less damage, and — this is the part that decides fights — you are spending turn slots you do not have. You only play three cards a turn. A card that costs 3 SP and deals 250% is worse than a card that costs 1 SP and deals 270%, and it is worse by more than the numbers say, because the expensive one clogs your hand on turns when the pool is empty.
The proven endgame shape, from a player at chapter 16:
| Hero | Role | Core cards |
|---|---|---|
| Copernica | Damage | 3-4× Fire Pillar, plus 3-5 one-cost action cards (Arcane Detective, Burning Desire, Searing Lash) |
| Armilly | Fire support and interruption | 2× Intercept, 1× Fiery Bash, 1× Encourage, 4× Hot Iron |
| Orik | Buffs and control | 2× Stone Lion (0 cost, guaranteed intercept), 2× Bushido, 2× Iai, 1× Fascination, 1× Inspiring Grace |
Bushido, upgraded once, gives 75% crit chance to any character — the single biggest damage multiplier in the party, and it is a buff, not an attack. Stone Lion costs zero and guarantees an intercept. Orik’s weapon ability lets you play more cards per turn, which is the only resource that actually matters.
And the honest summary from that player: “I almost don’t use any cards costing more than 1.”
Why the deck maths forces this
Each hero has eight card slots. Three heroes make a 24-card battle deck. You hold six and draw three per turn.
That produces a rule that is easy to verify and hard to internalise: any card not currently in your hand has about a one-in-six chance of appearing each round. You are not going to dig for a specific expensive card. You are going to draw whatever the deck is made of, so the deck has to be made of things that are always playable.
Which gives the benchmark:
- Skill cards (SP consumers): 7 to 10 out of 24. Never above half. Players who fill the deck with skills report hands that “look impressive and do very little.”
- Everything else: Strikes and Upgrades — the cards that build Steam Pressure so the skills can fire.
Steam Pressure is a party resource, not a per-hero one. That single fact reframes the whole game: the question is never “what is the biggest hit I can land,” it is “what does my next hero need.”
Chains, and the one rule that ruins them
Playing three cards from the same hero in a row triggers that hero’s Chain — extra damage, defence, or an effect depending on the weapon.
The trap is specific to Armilly. Her Valiant Sword chain attacks the enemy you last attacked — and if that enemy is already down, it hits a random target.
So your final attack before the chain must be non-lethal, or you lose control of where the chain lands. That is a genuinely counterintuitive instruction: deliberately not kill something, so your bonus does not waste itself on the wrong body.
The general principle: do not force a three-card line just for the chain. If two plays are close in value, take the one that also gives a chain. If they are not close, take the better play.
Redraws are free, and that is the point
Two redraws per turn, no penalty, discarded cards cycle straight back into the deck. Copernica’s Focus card grants an extra redraw temporarily.
The mistake is redrawing before you know your plan. The correct order:
- Read the enemy side. Is one nearly dead? Is the boss about to act?
- Decide the three-card line you want this turn.
- Then redraw the cards that do not serve that line.
Redrawing to fish for a miracle is how you burn two redraws and still play a bad turn.
Three habits that decide more fights than card quality
Heal before the final blow. The battle ends the instant the last enemy dies, and anything queued after the killing hit never resolves. If you are low and about to win, heal first, then finish. This matters most on hard and before long stretches without a save point.
Focus one enemy down. A dead enemy stops acting. Multi-target damage looks efficient and usually is not — fewer attackers means fewer status effects, less chip damage and less healing pressure. Save area damage for when several are already low.
Exploit the last enemy. Down to one enemy with decent defence and a heal, you can stop attacking and heal the party back up while it plinks at you. Slow, safe, and better than arriving at the next room at half health.
Early-game errors worth fixing in the first hour
- Remove Book Bash from Copernica immediately. It is a physical attack on the party’s lowest-Strength character. It does 7 damage. Let Armilly and Galleo handle physical and let Copernica build Steam with Mana Barrier.
- Do not use one copy of a card and expect never to see it. Single copies come up plenty — the deck maths above says so. But do not run three copies of a situational card either.
- One card copy is usually enough. Eight slots per hero is tight; duplicates belong on your basic attack, not on your finisher.
- Consumables are a last resort, not a plan. Carry roughly six Recovery Vials, three Ointments and three Repair Kits, but prioritise card healing — items do not contribute to chains. The exception is the Repair Kit, the only mid-battle revive. If you can win without the downed hero, do — they auto-revive at 1 HP after the fight and you save the kit.
- Dying costs almost nothing. You go back to the last Hero Statue, fully healed, keeping XP, gold and treasure. Experiment freely; there is no build you can ruin.
Related guides
- SteamWorld Quest guide — the campaign structure
- Quest chapters and bosses — where each fight sits
- Quest all cards — the full card pool
- SteamWorld games ranked — where Quest sits in the series
- Series order guide — what connects to what
- Beginner guide — series-wide starting points