There is a moment in almost every game of Yahtzee where the dice have stopped cooperating, the useful boxes are full, and you have to write a zero somewhere. Most players treat this as a failure to be minimised. It is actually a decision, it comes up three or four times a game, and handling it well is worth more than any single lucky roll.
Zeros are inevitable, so plan for them
Thirteen categories, thirteen turns, no skipping. If a turn ends with nothing that fits, something gets a zero. The only question is which.
The useful reframing: a zero is not a loss of points, it is a loss of the difference between what that box could have held and nothing. Zeroing ones costs you at most five points and realistically about three. Zeroing the large straight costs forty. Those are wildly different events described by the same word.
So the first principle is to keep a cheap box open as long as possible. Ones is the standard sacrifice - low value, and a short score there barely dents the upper bonus budget.
The sacrifice order
Roughly, from cheapest to most expensive to zero:
| Category | Typical cost of a zero | Notes |
|---|---|---|
| Ones | ~3 | The default dump. Keep it open early. |
| Twos | ~6 | Second choice, but watch the bonus budget. |
| Yahtzee | ~50, but only 4.6% likely | Cheaper than it looks - most games have no Yahtzee |
| Threes / Fours | 9-12 | Now the bonus is at risk |
| Full house | 25 | Reasonably likely to fill; do not zero early |
| Small straight | 30 | Quite likely to fill |
| Large straight | 40 | Expensive, but hard to fill |
| Fives / Sixes | 15-18 plus the bonus | Nearly always the worst zero on the sheet |
The Yahtzee box is the interesting one. Because you will roll a Yahtzee in fewer than one game in four with average luck, holding the box open "just in case" all game is usually a mistake - it means you are sacrificing elsewhere to protect a lottery ticket. Zeroing it around turn ten, if nothing has come, is frequently correct.
Chance is a pressure valve
Chance - score the sum of all five dice, no requirements - is the most misused box in the game.
Players spend it early on a mediocre roll because it is easy, and then have nowhere to put a bad turn at the end. Chance is the only genuinely flexible box on the sheet, which makes it worth far more late than early.
The guidance: hold Chance until roughly turn ten, and do not use it for less than about 20. A Chance of 14 is close to a zero with extra steps; a Chance of 26 is a solid score you could not have placed anywhere else.
Holding a flexible resource rather than spending it on the first convenient opportunity is one of those principles that shows up everywhere. It is exactly what the flowers and seasons are for in tile matching - take Mahjong online: the bonus tiles match any other bonus tile, which makes them your escape hatch when the board tightens, and cashing them early for a small satisfaction is how a winnable board becomes a dead one.
The three-roll decision tree
Most zeros are avoidable with better keeping decisions two rolls earlier. A quick framework:
After roll one, ask what the roll is closest to, not what it is. Four dice within a straight range is a straight draw, even if the pair looks tempting. Three of anything is a three-of-a-kind draw.
After roll two, commit. Splitting your keeps between two objectives - holding a pair and a straight draw simultaneously - is the classic error, because a single re-roll rarely completes either. Pick the target with more outs.
Count outs, not hopes. An open-ended straight draw needs one of two numbers on one die: roughly a one-in-three chance per roll. A pair going for three of a kind on three dice is better than it feels. Four of a kind going for five is one in six per roll.
This is the same accounting that every game with bounded randomness rewards - counting what can still arrive rather than hoping. In a game of Tetris, the pieces come from a shuffled bag containing one of each, so a strong player knows precisely how far away the piece they need can possibly be. Yahtzee's dice have no memory, but the per-roll probabilities are just as countable.
Endgame: match the sheet, not the dice
In the last three or four turns the calculation changes. You are no longer building a scorecard; you are filling specific holes.
Look at what is open before you roll, and keep dice that serve those boxes even when a better-looking combination is available. Holding three sixes is pointless if the sixes box is full and you need a small straight.
That sounds obvious and is routinely violated, because the dice are in front of you and the sheet is off to the side. Deliberately checking the sheet before each keep decision is a small habit with a real payoff.
It is the same discipline as playing to the position rather than the pieces in a game with no hidden information at all - in a game of Reversi, the attractive move that flips the most discs is usually the wrong one, and the correct move is chosen by what the board will need later.
How the variants differ
Yatzy adds one-pair and two-pair categories, which gives you two more cheap boxes and makes zeros substantially rarer - it is a gentler game for exactly that reason. Maxi Yatzy uses six dice and more categories, so the allocation problem gets harder rather than easier. Triple Yahtzee runs three cards with multipliers, which turns the zero decision into the central skill of the game - you are choosing which card to damage.
Generala and Balut restructure things further, and Kismet adds colour-based categories that reward a different kind of keeping.
Try this
Play one game in which you refuse to use Chance before turn ten and refuse to score any upper category with fewer than three of that number. You will take two or three zeros in ones and twos and will probably score better than usual, which is a useful thing to see happen.
The rules hub covers every variant's categories, the scoring FAQ gives realistic benchmarks, the luck-versus-skill FAQ is a good companion to this piece, and the glossary names the terms. Practise on the main game.
For the upper-section bonus that most of these decisions revolve around, see the companion piece.