Dominoes Strategy, Measured

How these numbers were produced

Every figure on this page comes from 15,000 simulated rounds per experiment on a double-six set, seed 20260805, run through the exact engine that plays the game above. The script is scripts/run-sim.mjs in the site’s source, seat order is alternated to cancel the opening advantage, and the seed is fixed so anyone can reproduce the figures rather than take our word for them. The simulations play the single-line base rules — no spinner — because those are the rules these pages document; the game above offers the spinner variant as well.

One limitation stated plainly: these are results between computer opponents, not between people. They measure what the described strategies are worth against each other under the rules as implemented. They are evidence about the game, not about your family.

What is leading actually worth?

"Whoever holds the highest double leads" is the rule everywhere, and it is usually described as an advantage without a number attached. Here is the number.

Leading is worth about four points of win rate heads-up — real, but not decisive. At a four-player table it is worth nearly eight, almost double. That is the opposite of the intuition that a crowded table dilutes any single player's edge.

The mechanism is tempo: with four hands to get through, the opener is consistently a move ahead of three opponents rather than one, and rounds end sooner relative to the number of tiles each player holds. It also means the “highest double leads” rule hands out a bigger advantage the more people are playing — worth knowing if your group plays a long series and never rotates who deals.

A related figure: somebody is dealt a double in 99.7% of deals, so the "heaviest tile leads if nobody has a double" fallback almost never fires. In a thousand deals you will see it about three times.

How often rounds end blocked

This is the most practically useful number on the page. In the Block game, 61% of rounds jam — so the blocked-game scoring rule is not a footnote, it is how you win or lose most rounds. If your group plays Block and has never settled whether the winner subtracts their own pips, you have an unresolved argument sitting under the majority of your scorepad.

In the Draw game it collapses to 19.2%, because a stuck player refills instead of stalling. Adding more players lowers it further still (15.6% at four), since more hands means more chances that somebody can answer the open ends.

The heavy-hand advice is wrong

"Get rid of your heavy tiles early" is the most repeated tip in dominoes. Tested, it does not hold up the way it is usually stated.

The blocked rate barely moves with hand weight — 18.7% to 20.3% across the whole range. And the players dealt the heaviest hands finish holding slightly FEWER pips (9.3) than those dealt the lightest (10.9).

That looks backwards until you think about what a light hand is. A light hand is full of low tiles and blanks, which match fewer numbers and give you less to do — so you sit on them. A heavy hand is full of tiles you are motivated to spend and that fit the layout, so you spend them. The advice is not wrong about wanting to shed weight; it is wrong about a heavy deal being bad news.

The practical version: your starting weight is not the thing to worry about. The number of DIFFERENT numbers you hold is. A hand of seven tiles spread across six numbers can always answer something; a hand of seven tiles concentrated in three numbers is the one that gets stuck.

What actually wins

These numbers are a useful answer to "is dominoes a game of luck?". A player who counts unseen tiles and tracks what opponents have passed on beats a competent but unreflective player 55.6% of the time — a real edge, but a small one. Against a player who simply plays something legal, it rises to 63.7%.

So: dominoes is mostly luck in any single round, and skill is decisive over a long match. About five points of edge per round compounds into near-certainty across a race to 100.

• Count what is still out. Every number appears on seven tiles in a double-six set. If five of the 4s are visible between your hand and the table, an end showing 4 is very likely to stick your opponent.
• Remember every pass. A player who passes with 3 and 5 showing holds no 3s and no 5s — for the rest of the round. That information never expires and it is free.
• Play doubles early. A double matches one number and one number only, so it is the hardest tile to place and the worst to be caught with.
• Watch for both ends showing the same number. If you hold several of that number it is a commanding position; if you hold none, you are one bad turn from passing on your own layout.
• Count the boneyard, not just your hand. In the Draw game the last two tiles are never drawn, so the moment the stock hits two, the game turns into Block and passing becomes possible again.

Strategy questions

Is dominoes a game of skill or luck?

Both, in a measurable ratio. Across 15,000 simulated rounds, a player who counts unseen tiles and tracks opponents' passes beats a competent player only 55.6% of the time — so any single round is mostly luck. Over a race to 100 points that edge compounds, which is why the same people keep winning domino nights.

Should you play your heaviest dominoes first?

Mostly yes, but not for the usual reason. Heavy tiles cost you if the round blocks, so shedding them is sensible — but our simulation found that players dealt heavy hands actually end up holding fewer pips than players dealt light ones. Your real risk is holding tiles concentrated in a few numbers, not holding high numbers.

What is the best opening move in dominoes?

You rarely get a choice — the highest double leads, and in 99.7% of deals somebody holds one. When you do have a choice of tiles later, favour the ones that leave an end you hold several of, and avoid spending your only tile in a number you are about to leave open.

Does going first matter in dominoes?

Yes, and more than you would expect at a big table. Heads-up, the opener wins 54.3% of rounds against a 50% baseline. At a four-player table the opener wins 32.6% against a 25% baseline — an edge of nearly eight points, almost double the two-player figure.