How can you predict what your opponent will play in the next 5 turns, 7 turns, 10 turns?

If the game is like a tree, then the more the turns the more it will branch out.

How can a player accurately predict what their oppoent will play several moves into the future when the number of possible combinations of moves to get ten moves ahead can be ridiculously large?

Please give a detailed explanation.

How do I learn/acquire this skill?

  • It depends upon the level of opponent. Jan 27 '16 at 18:27

The answer is that you only consider the most sensible and most critical options. This prunes the tree down to manageable proportions.

One easy example would be the fact that retreats are usually not threatening and therefore less likely to be critical. (Of course the downside is that actually dangerous retreats are often overlooked.)

To learn how to recognise potentially critical moves you have to play through master games, solve puzzles and play games. Over time your brain will store typical moves and manoeuvres in certain position. This ability is called pattern recognition. The more patterns your brain stores the stronger its intuitive move suggestions will be, you will be able to prune the tree more.

Another aspect is called "forcing moves". A forcing move is a move that requires a certain response, i.e. severely restricts the opponent's choices. For example checks, threats of checkmate and other kinds of strong threats. If your moves are very forcing the tree will resemble a path, because every step from this path is either illegal or results quickly in very bad things. If GMs calculate very long variations, these usually contain very forcing moves. In positions without forcing moves even the best players don't look very far ahead.

There is also a non-intuitve conscious part to "seeing ahead". Even if your brain provides you with only the most likely good moves, you still have to visualise the outcome of every variation in your head. That requires training as well, for example solving studies or tactical puzzles with long variations.

  • 1
    However, this does give a dangerous risk of not seeing non-intuitive moves. Often the biggest mistakes come from missing these moves which are often followed by extreme cramping or a tactic.
    – Mildwood
    Jan 27 '16 at 23:27

I would recommend Dan Heisman for detailed explanations of how effective analysis is done in Chess.


Unfortunately, many of his articles have gone behind a membership wall now (at chesscafe.com), but not all:


To answer your question in short: you don't (predict 5,7,10 moves ahead); instead you find dangerous moves (of your opponent, then of your own) then look further ahead only on critical moves for stopping/minimizing the opponent's danger while maximizing your own danger.

(Note that in both of Dan's examples at the second link, you only need to "look ahead" at most 2 moves.)


Other answers are also good. But I would explain it simpler.

You just take account of few best moves, you don't care others because you know that they are inferior you can refute them easily. You keep calculating positions which are unclear for you, until you assess them.

For example, if something lets you win a queen in a middlegame position, you know that is almost always winning for you(you assess it winning), and you stop calculating.

Forcing moves let's you go deeper because if there are forcing moves it means there are less options(generally only one option, think about checks for example).

So to conclude, it's generally easy get 5 moves deep, some positions lets you easily calculate 30 moves ahead, it really depends complexity of the position, and experience you have. Crucial thing is, who is calculating deeper and who is selecting candidate moves more correctly, that's the thing that decides the winner in a game.

Think Like a Grandmaster by Kotov is an excellent book which covers thinking processes in chess. I strongly recommend you to read this book.

  • 3
    Umm, "easily calculate 30 moves ahead" is a pretty horrible exaggeration. Also, the third paragraph contradicts the (very good) maxim that "once you find a good move, always spend at least some time looking for a better one". (Assuming a normal (non-blitz) game.)
    – Jeff Y
    Jan 27 '16 at 22:53
  • You can easily calculate 30 moves ahead King+Rook vs King for example. It is not better to look a for better move, if you have a move which is winning already, unless you have a lot of time to waste.
    – ferit
    Jan 27 '16 at 22:59
  • And you can calculate infinitely deep perpetual check positions for example.
    – ferit
    Jan 27 '16 at 23:00
  • If you could "easily calculate" the moves of a King+Rook vs King ending, you would "easily" be able to say on which square the lone King will be checkmated. But you can't. (Let alone that K+R vs K should never take 30 moves to begin with.)
    – Jeff Y
    Jan 29 '16 at 12:35
  • 2
    Assessing and calculating are not the same thing. In the context of the original question, explaining the difference between them would help the OP more, instead of just intimating that assessing the position is the same as "predicting what your opponent will play".
    – Jeff Y
    Jan 29 '16 at 16:00

forced moves and pattern recognition simplify doing combinations

the risk is that some rare move that was overlooked may be better than the obvious 'forced' one or that the pattern has some tiny difference this time and so fails

but for a quick move selection it works well however for postal chess you might want to put more effort into ensuring the moves are truly forced (or at least best for the other side)


One way to see a lot of moves ahead is to play postal chess.

That is a variant that allows both players to work out multiple variations following each move.


Deep chess requires the player to filter out a set of plausible moves given an infinite series of permutations. The filtering, in effect, is dependent on both the player's strength and tactics.

Dependent on both time and position, considering the relative value of a branch down the tree of moves is typically calculated by a human as cumulative "face value" listing each move considered. The best position in a game, however, is most certainly NEVER the sequence of "best moves" looking 2-3 moves deep per turn, for example.

Truly strong players will consider most options beyond the obvious, yielding subtle and deep moves that support underlying tactics and combat human psychology. Check out Mikhael Tal's games -

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.