Mikhail Tals' books on chess. Learning/analytic chess theory. Perception and intuition in the development of strategic and tactical chess play.

  • 1
    Not exactly the same skill, but some people are able to play entire games blindfold. This entails an ability to maintain a mental board state for dozens of moves. If Tal had that ability, then he would be able to in effect play blindfold chess with himself, starting with the position on the board, for any number of moves. Some masters can even play simultaneous blindfold games. If Tal had that ability, then rather than just projecting a single game state into a hypothetical future he could in effect do several and keep a branching tree of game states in mind. May 8, 2020 at 1:36
  • 1
    The phrase "50 variations ahead" does not make much sense. Are you sure about this exact phrase? Could it be a translation error? More meaningful would be "50 moves ahead" or "50 variations to depth....". May 8, 2020 at 6:58
  • I claim to have spoken to god, even though I don't believe in your god. I believe this to be an exaggeration.
    – Mike Jones
    May 8, 2020 at 12:39
  • 50 moves ahead is obviously not humanly possible, but if you consider looking at 50 variations: Even if you are starting with only 5 moves (which I think is reasonable to look at in most positions), you can look at 50 variations in less three moves deep. Which sounds sensible, even for a non-world-champion, to do. May 9, 2020 at 12:22

3 Answers 3


Note that "number of moves ahead" is not necessarily a great proxy for "number of positions considered". An open board with no impending threats has many possibilities of what might be the "best" move, and will require a player to analyze a wide, but shallow tree of possibilities. A board with few pieces and many threats has relatively few "good" (or even legal) moves, so a player can look farther ahead down a narrow but deep tree of possibilities. In some cases, the tree becomes a path - there is only one move to make. This question, for example, shows games with many repeated checks, making the tree of potential moves very narrow indeed.

It seems that most high-level chess players tend to think 20-40 moves ahead, depending on the board state. It's not unreasonable to think that someone could think 50 moves ahead in very particular situations, where the next few moves are "obvious" or even forced. But to be able to consider 50 moves ahead in all possible situations is likely beyond a human's ability.

  • 3
    The OP did not say "50 moves ahead". He said "50 variations ahead". I don't know what that even means. I wonder if OP is quoting (or translating) Tal accurately.
    – bof
    May 7, 2020 at 22:27
  • @bof I don't know what "variations ahead" mean either, so I went with the only reasonably interpretation I could come up with. What else does someone "see ahead" in chess other than moves? May 8, 2020 at 13:02

"50 variations"

This is very possible. I think that the key to understanding this is to know the difference between variations and moves ahead.

VARIATIONS: Consider any move. What can the opponent do in return? Does he have 2 choices? 30? Each of his choices makes a variation. If you have a candidate move and you opponent can reply in 10 possible ways, then you see 10 variations.

MOVES AHEAD: I play e4. Then I think that you will play e5, I Nf3, you Nc6, I Nc3, you Nf6, I Nxe5, etc. (Halloween Gambit). This involves thinking 6 moves ahead on a single variation.

So 50 variations? I think that you can do that. In most positions you have about 3 candidate moves. Consider each move; think about three or four moves ahead for each one. You will certainly run into variations. To make 50 variations, your three candidates will have to have about 16 variations each. Each candidate is a variation of of the current position. In our case of 3 candidates, each candidate will demand 4 variations which have 4 of 5 variations each to make 50 variations (4x4x3 = 48). So you see, you variations add up rather quickly. I think that to make 50 variations, you would have to see only about 4 moves ahead in most positions.

  • Adding up variations this way is pretty trivial. There are 20 opening moves in a chess game, and your opponent can respond with any of their own 20 possible opening moves. There are 400 variations in the first two moves alone, and you're in the hundreds of thousands by the next two moves. May 11, 2020 at 18:53
  • @NuclearWang I agree. Although most variations in any given position can be discarded as blunders. I would not count these as calculations. I do think that as the game advances (in terms of time and player skill) the variations that actually count become more difficult to calculate. The advanced player often has fewer candidate moves with more variations, whereas the novice often has more candidates and fewer variations. Actually counting them, however, is rather pointless.
    – Senrab
    May 11, 2020 at 20:34

In my opinion, this is totally possible, as it has been said above good players have the ability to play games blindfolded, and players such as Tahl are skilled enough to play several blindfolded games at the same time. But in a practical way, calculating 50 moves ahead doesn't make any sense as the possible variations for 50 moves are infinite, even though good players have the ability to go this deep in a concrete line, they don't have to because their instinct will tell them much before whether that's a good line without having to go that deep.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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