Casino War: Go to War or Surrender? (Always Go to War)

Casino War has exactly one decision, and it only comes up on a tie: go to war, or surrender. It feels like a real choice, but it isn’t — the math says go to war every single time. Warring is a 2.88% house edge; surrendering every tie is worse at 3.70%. Here’s why, and why there are no exceptions.

The only decision in the game

The base game is a coin flip: you and the dealer each get one card, higher wins even money. About 7.4% of the time (roughly 1 in 13.5, with 6 decks), the cards tie — and that’s the one moment you have a choice:

• Go to war. You raise a second bet equal to your ante. The dealer burns three cards and deals one more to each of you. If your card is higher or equal, you win; if it’s lower, you lose.
• Surrender. You give up half your ante and the hand ends.

Why war beats surrender

Surrendering feels safe — you cap the damage at half a bet instead of doubling down on a coin flip. But “safe” and “cheaper” aren’t the same thing. Handing back half your ante on every tie is a steady, guaranteed leak. Work it out over all ties and surrendering costs 3.70% of your ante.

Going to war costs less — 2.88% — even though it risks more per tie, because you actually win most wars. Ties in the war go to the player, so your card wins whenever it’s higher or equal. You come out ahead on a bit more than half of wars, and the even-money win on the raise recovers far more than surrendering’s guaranteed half-loss.

The rule: always go to war. Never surrender. It’s not situational — there’s no card, streak, or read that flips it.

But war still favors the house — here’s the catch

If you win most wars, why does the house still keep 2.88%? Because of an asymmetry in what you risk versus what you win. Going to war puts two units at stake (ante + raise), but a win only pays one: the raise pays even money, while your original ante merely pushes (you get it back, no profit). So a war win nets +1, a war loss costs −2. Winning a little over half the time isn’t enough to overcome risking two to win one — and that shortfall is the entire 2.88%.

Surrendering tries to dodge that asymmetry by bailing out, but the half-ante toll adds up to more than the war math costs you. The lesser evil wins.

No exceptions, no counting

Casino War has no strategy beyond this. There’s no basic-strategy chart, no card counting worth the effort, nothing to memorize. Every tie, every time: raise and go to war. The only other “decisions” on the table — the 10:1 Tie side bet and hunting for favorable rules and deck counts — happen before the cards come out.

See it for yourself: Play Casino War free →, go to war on every tie, and watch the observed edge settle toward 2.88% in the Lab — always below the 3.70% you’d bleed by surrendering.