Dominoes strategy: how to actually win

Dominoes looks like a game of luck for about three rounds, and then it stops. The deal decides what you are holding; almost everything after that is a sequence of decisions about which end to leave open and which weight to get rid of first. What follows is the reasoning that separates a player who places whatever fits from one who is steering the board.

1. Go out if you can. Nothing else is close.

Playing your last tile ends the round immediately and hands you every pip your opponent is still holding. No positional advantage is worth passing that up. It sounds obvious, and yet the most common real mistake in casual play is making a clever blocking move while holding a tile that would have ended the round.

The corollary matters too: once you are down to two or three tiles, start thinking about the sequence that gets you out, not the sequence that looks strongest. Count whether both of your remaining tiles can actually be played on the ends you are about to create.

2. Shed weight early

Every pip in your hand at the end of a round is a point for somebody else. A typical seven-tile hand is worth around forty-two points, and if you get caught holding [6|6], [6|5] and [5|4] you have handed over thirty-three of them in three tiles.

Play heavy tiles while you still have turns left to place them. Early in a hand you have the most open options and the least information, which is exactly when a heavy tile is cheapest to unload. Late in a hand, both ends are constrained and a [6|6] can become unplayable.

3. Play your doubles early

A double is the least flexible tile in the set. [4|4] joins only fours, and it exposes only a four. Every other tile gives you two different numbers to work with. Holding a double into the endgame is the most reliable way to get stranded, because you need a specific number to still be open at the exact moment it is your turn.

This is why the coach gives a bonus to playing a double that decays as your hand shrinks: early, you have turns to spare and the double is a liability worth clearing. Late, the bonus no longer matters because the decision has usually been made for you.

4. Keep your hand broad

Count the distinct numbers in your hand, not just the tiles. A hand of six tiles spanning five different numbers can answer far more boards than six tiles spanning three. When two moves look equally good on pip count, take the one that leaves you covering more numbers.

The reverse is the classic trap: dumping tiles so efficiently that you end up holding three tiles that all show the same two numbers. You have a light hand and no way to play it, which in Block means passing repeatedly while your opponent walks out.

5. Control an end you can answer

When you place a tile, you choose what number the new open end shows. Choose a number you hold more of. If you leave a five showing and you still have two more fives in hand, you have effectively guaranteed yourself a move next turn, no matter what your opponent does at the other end.

The inverse is the mistake worth naming, because it is the one that quietly loses rounds: opening an end in a number you hold none of, in a suit that is nearly played out. Seven tiles carry each number. If five of the threes are already on the table and you hold none of the rest, leaving a three showing means you are relying on your opponent to rescue you.

Counting a suit

You can do this properly rather than by feel. Seven tiles show each number. Count how many are visible on the table, add how many you hold, and subtract from seven: that is how many could possibly answer that end, whether they are in your opponent's hand or the boneyard. When the answer is zero or one, that end is nearly dead and whoever cannot play on it is in trouble.

6. In the endgame, block

Once your opponent is down to two or three tiles, your priority shifts. Stop optimising your own hand and start setting ends they cannot play. And you often know exactly which ones those are, because they told you.

A pass is permanent information. Passing is only legal when no draw is available, so a player who passes on a board showing a two and a five holds no twos and no fives, and can never acquire one for the rest of that round. Three passes into a hand you may know four or five numbers they are void in. Set one of those and they pass again while you play on.

A draw is weaker evidence: they were stuck at that moment, but the tile they took might have been the one they needed. Worth noticing, not worth relying on. The app shows you the hard evidence, the numbers your opponent has actually passed on, above their hand, so a first-time player is not at a disadvantage against someone who tracks it in their head.

If both open ends are numbers your opponent has passed on, and they have two tiles left, you have effectively won the round. They cannot move, and every turn they pass is a turn you spend emptying your hand.

Where the two variants diverge

In Block, information is everything: no tiles enter a hand, so every pass narrows the possibilities permanently and blocking is devastating. In Draw, the boneyard softens all of it. Blocking is less final, bad hands can be repaired, and raw pip management matters more relative to inference.

One practical adjustment: in Draw, be more willing to make a slightly awkward heavy play, because if it strands you there is a boneyard to fish in. In Block, that same play can end with you passing four turns in a row.

Practising against the coach

Tap Analyse board during your turn. The coach names the move it would make, ranks the reasons behind it as bars, and states the trade-off you are accepting. Try forming your own view first, then check it. Where you disagree with it is where you will learn something, and it is not always right: it is a heuristic, not a solver, and it does not search ahead.