Math in Chess: Counting Moves and Thinking Ahead

My son learned the knight first, because it jumps. Then he asked the question that stops most chess conversations: how many different games of chess are there? Not how many he would play in his life. All of them.

He expected a large number, the way children expect dinosaurs to be big. What he got was a number that broke his sense of scale completely, and took the rest of the evening to explain. The short version is that chess is not a game of hidden cleverness so much as a counting problem with more branches than the universe has atoms. And once you see why the counting fails, the phrase “thinking ahead” stops meaning what it sounds like.

The first move is small enough to check

Start with what you can verify on a real board, because the whole article depends on trusting the numbers later. Set up the pieces and count White's legal first moves. The eight pawns can each go one or two squares, which is sixteen options. The two knights can each reach two squares, which is four more. Nothing else is free yet.

Sixteen plus four is twenty. That is the entire second row from the bottom of the famous table, and your child can confirm it without taking your word for anything. This is the most valuable move in the whole lesson: establish that the counting starts somewhere ordinary and checkable, before it becomes unimaginable.

Then it stops being countable by hand

Now let both sides move. After White's move, Black has roughly twenty replies, so two half-moves produce about four hundred games. Three half-moves gives 8,902. Four gives 197,281. These are exact counts, not estimates, and they were worked out by careful enumeration.

Look at the last row honestly. Ten half-moves is five moves each — roughly the part of a game a beginner is still learning. At that point there are already sixty-nine trillion distinct games on the board, and no game has been decided. The counts are tallied in the reference account of the Shannon number, along with the growth that follows.

The average number of legal moves at any position, called the branching factor, is about thirty-five. That single number is enough to explain the rest, because multiplying by thirty-five over and over is what turns twenty into sixty-nine trillion in ten steps. This is the same compounding that our exponents guide covers in the abstract, except here you can feel the size arriving.

The man who did this arithmetic on purpose

In 1950 Claude Shannon wrote a paper called “Programming a Computer for Playing Chess,” and its purpose is often misunderstood. He was not showing off how clever a machine could be. He was working out whether a machine could simply try everything, and his conclusion was that it cannot.

His calculation: about a thousand possibilities for each pair of moves, times about forty such pairs in a typical game, gives roughly ten to the hundred and twentieth distinct games. That is a conservative lower bound, not an estimate, and it is the figure known as the Shannon number. Compare it with the number of atoms in the observable universe, which is usually placed around ten to the eightieth, and the gap is about forty orders of magnitude. Not “more than you would think.” More than every atom, by a margin so wide the atom count barely registers.

A computer that examined a billion positions a second would still be nowhere near finishing, and would be nowhere near it after the sun burned out. Shannon's paper is usually remembered as the beginning of computer chess. It reads more like a proof that the obvious approach was dead on arrival.

So what does “thinking ahead” actually mean?

This is the payoff, and it reframes the whole game. If you cannot examine everything, then skill is not the ability to look further. Skill is the ability to decide which few branches deserve looking at and which can be thrown away without regret. Every strong player is doing pruning, not searching. They are not calculating more positions than you. They are discarding better.

Two questions do most of that discarding. Does this move force an answer — a check, a capture, a threat that cannot be ignored? And what is the opponent threatening if I do nothing at all? A move that fails both tests can usually be set aside without analysis, which is how a player holds a thirty-five-way branching position in a head that fits one plan at a time. Beginners do not lose because they look too shallowly. They lose because they try to look everywhere at once and run out of time before reaching the interesting part.

The same idea runs through the counting shortcuts we teach in school. You never list all the possibilities either; you multiply the choices and skip the listing, which is exactly what the multiplication principle guide is for. Chess applies that instinct under a deadline. The board is asking the same question a good counting problem asks — how many branches are there, and which ones do I not need?

There is a cleaner way to picture it. A probability tree draws every branch and multiplies along the path. Chess is the same tree, wide enough that drawing it is impossible, so the player chooses a thin slice in their head and hopes the slice contains the truth. That hope is what experience is made of.

Two things this number does not mean

The first mistake is concluding that chess is therefore unsolvable and its outcomes mysterious. It does not follow. The size of the tree tells you brute force is impossible; it says nothing about whether a perfect answer exists. Strong machines have since shown that human play is far from optimal, and they manage it precisely by pruning rather than by counting everything.

The second mistake is treating the number as a fact about skill. The ten-to-the-hundred-and-twentieth figure is a lower bound on games, not a measure of difficulty, and two players who both know it are not equally good. The number explains why the game is hard to exhaust. It never explains why one player wins, which is a different question entirely.

Worth keeping separate from the strategy layer, too. Our game theory guide covers what happens when players reason about each other over repeated meetings. This article is only about the size of the tree they are standing in. The counting is the floor, not the game.

Count the first move at home

Set up the board and do the twenty-move count together, using your finger on each square. Then ask your child to predict the count after two half-moves before checking it against four hundred. Most will guess something like forty, because the mind expects addition when it should expect multiplication.

That wrong guess is the lesson, and it is worth naming out loud. Twenty times twenty is four hundred, not forty. When something grows by multiplying instead of adding, intuition underestimates it every time, and chess is where the underestimation becomes spectacular. Our scientific notation guide then gives them the tool to write such numbers down without counting zeros for ten minutes.

Write the four numbers down as you go — 20, 400, 8,902, 197,281. The shape of that growth is visible without explanation, and a child who has written it down is far more willing to believe the later sentence about ten to the hundred and twentieth.