Battleship Strategy: The Probability Method That Wins More Games

Here is a number worth knowing before your next game: a data scientist named Nick Berry once simulated hundreds of millions of Battleship games to find the best possible way to fire. Ordinary "guess, then chase the hit" play clears an opponent's fleet in about 66 shots on average. His best method does it in a median of about 42. On a 100-cell board, that gap is the difference between usually winning and usually losing.

You don't need a computer to steal most of that edge. The whole trick is learning to see where ships are likely to be instead of firing wherever your eye lands. Below is the full ladder — from the way most people play up to the "hot zone" method — plus the defense that most guides get exactly backwards.

The rules, in one breath

Neon Parlor's Battleship game uses the classic setup: a 10×10 grid for each player and five ships — a Carrier (5 cells), Battleship (4), Cruiser (3), Submarine (3), and Destroyer (2). You place them, then take turns calling shots at the opponent's grid. Hit every cell of a ship to sink it; sink all five to win. One detail matters for strategy: in this game ships may sit directly next to each other — there's no forced gap between them. Remember that; we'll weaponize it later.

Level 1: Hunt, then Target

Almost everyone already does a version of this without naming it. You're in one of two modes at all times:

• Hunt mode — no active leads, so you're searching for any ship.
• Target mode — you just scored a hit, so you stop searching and fire at the four cells touching it until the ship sinks. Then you drop back to Hunt.

Simply committing to this — never wasting a shot searching while a wounded ship sits half-sunk — is what gets you to that ~66-shot baseline. If you're not doing it deliberately, doing it deliberately is your single biggest upgrade. It's the same discipline that separates casual and competitive play in any turn game, from checkers to 8-ball pool: finish what you start before you go looking for the next thing.

Level 2: Parity — useful, but not the miracle it's sold as

Here's a tip you'll see in almost every Battleship guide: treat the board like a checkerboard and only fire at, say, the dark squares. The logic is sound. The smallest ship, the Destroyer, is two cells long, so it must cover one dark and one light square. Fire only at dark squares and you're guaranteed to eventually clip every ship — while searching half as many cells.

That's real, and it's worth doing. But be honest about what it buys you. In Berry's simulations, pure parity trimmed the average from about 66 shots to about 65 — a rounding error. (Slate, 2012, on Nick Berry's analysis.) Why so little? Because parity only speeds up the hunting half of the game. Once you land a hit, you're forced onto the opposite color to sink the ship anyway. Parity is a good habit — it tells you which empty cells are worth searching — but if a guide sells the checkerboard as the secret to winning, it's overselling a warm-up move as the main event.

Level 3: Think in hot zones (the real lever)

The method that actually gets you to ~42 shots is a probability map. The computer version, recomputed after every single shot, works like this: for each unknown cell, count how many different ways the remaining un-sunk ships could still be placed so that one of them covers that cell. Cells that could hold ships in many ways are "hot." Fire at the hottest cell. That's it — but no human can tally thousands of placements per turn.

You don't have to. The map produces four instincts you can carry in your head:

• Search the middle first. A long ship has far more room to fit through the center than in a corner, where it's boxed in by edges. Center cells are hot; corners are cold. Open your hunt in the middle and work outward.
• After a hit, commit to the line. Two hits in a row mean the ship runs that direction — keep firing along it before you try anything sideways. This is just Target mode with conviction.
• Don't fire into a gap that's too small. If the biggest un-sunk ship is 3 cells and you've boxed a stretch of open water down to two empty cells, nothing can fit there. Those cells are dead — skip them. This is the single most-missed shot-saver.
• Chase the big ship. Early on, hunt as if you're looking for the 5-cell Carrier — its length means long open runs are where the probability piles up. As the big ships sink, the map cools and spreads out.

Flip the map: how to actually place your ships

This is where most guides contradict themselves. They'll teach you a center-weighted attack, then tell you to "spread your ships across the board and mix orientations." But if you know a good opponent is hunting the hot center, the best defense is to hide in their cold zones.

• Favor the edges and corners. They're the last places a probability map looks. A ship flat against the border sits in low-density water.
• Use adjacency — it's allowed here. Because Neon Parlor lets ships touch, you can cluster two ships side by side. Opponents are trained to assume open water around a sunken ship; parking a second ship right next to the first quietly breaks that assumption and wastes their follow-up shots.
• Mix your orientations so that discovering one ship is horizontal tells them nothing about the rest.
• Then vary it every game. The one thing worse than a predictable placement is the same predictable placement twice — a human opponent will read you. Edge-hugging is a strong default, not a rule to marry. (Placement trade-offs studied in Yale CSEC's edge-and-adjacency analysis.)

The fastest way to learn this: play a lot of it

None of this sticks by reading. It sticks when you watch a center-first hunt find the Carrier three turns quicker than random poking did last game. The cheap way to build that instinct is against bots that never rush you or rage-quit.

Once the method feels natural, the real test is another human who's also hiding on the edges and clustering ships. That's the game that actually moves — the mind game on top of the math.

Frequently asked questions

Is the checkerboard (parity) trick actually worth it?

Yes, but as a habit, not a headline. It tells you which empty cells are worth searching, cutting your hunt space in half. It barely changes the total number of shots on its own, though — the real speed comes from searching the hot center and not wasting shots on gaps too small for any remaining ship.

Can I really run the probability method in my head?

Not the exact math — that's a computer's job. But the four instincts it produces (search the middle, commit to the line after a hit, skip gaps that are too small, chase the biggest ship) capture most of the benefit with zero arithmetic.

Where should I hide my ships?

Against a strong attacker, favor the edges and corners and use adjacency to park ships side by side, since this game allows touching ships. Then vary it game to game so a human can't learn your pattern.

Is there a way to win every time?

No — placement is hidden and the opening is genuinely random, so luck always has a vote. These methods tilt the odds firmly in your favor over many games; they don't rig a single one.

Your next move: open a bot game, fire your first ten shots only in the center on one checkerboard color, and skip any gap too short for a 3-cell ship. You'll feel the difference by the second fleet.