In classical play, do GMs tend to resign in endgames like king+queen vs king+rook where the game can be won by the side with more material in under 50 moves, but only with good play? (I understand queen vs rook is a win in 31 in worst case, barring any immediate captures/checks which make it a draw. A computer can likely draw it out to 50 moves against a human, but another human might not.)

Queen vs rook isn't the only example - think of any pawnless endgame where the fastest forced mate is between 20 and 50 moves long.

On one hand I could see that it might be insulting to continue to play on with such a difference in material, where your opponent will clearly not fall for any skewer tactics and the endgame will be very long and very boring, but on the other hand I see the pragmatism in continuing play as long as there is a non-trivial chance of a draw.

  • If there are tricks, they will play. Just like RB vs R endgames are a draw, but actually very often end unpeacefully. Q vs R most also will try. The worst that can happen happens immediately if you resign.
    – B.Swan
    Commented Aug 20, 2020 at 19:34
  • They will fight while there's hope. If they know a computer can defend it against a human, they will also try themselves
    – David
    Commented Aug 20, 2020 at 22:01

1 Answer 1


They may as well try, since Q vs R is extremely difficult to win (even disregarding skewer tactics). The example of RB vs R is a great similar example, as it is a theoretical draw, but extraordinarily difficult to draw. There are numerous examples of top grandmasters being unable to convert Q vs R. As to a computer vs human- this was tested in 1978 in a special match between Belle (computer program, had the tablebase so it would play perfectly), and GM Walter Browne. Two positions were played, both with a win in 31 moves. He was unable to convert one but won the other (although barely avoiding the 50 move rule)

Example 1 - Gelfand-Svidler (two of the greatest players of their time, admittedly a rapid game. Black was pushing but Gelfand held.)

Example 2 - Morozevich-Jakovenko (Morozevich was a great player, but Jakovenko not so much. Morozevich was pushing, but Jakovenko held with what is known as a rambling rook idea, where a rook is sacrificed and keeps checking so either perpetual check occurs, or the king takes the rook leading to stalemate)

To add more to this- it's quite possible that the side with the queen is in time trouble, which heightens the difficulty of breaking through.

  • 1
    Note that the tablebase is playing the move that resists the most, not the move that makes it the hardest, so in practice it could even be more difficult
    – David
    Commented Aug 20, 2020 at 22:02
  • Eh, id argue that often they'd be the same, but I can certainly see the possibility. To make it clear as what you mean by hardest and resists: A tablebase will play the move, that assuming perfect play from both sides, will lead to mate (or conversion into a winning endgame, in this case, winning the rook) in the most (or least, if it's pressing) number of moves. Making it hardest for a human player is subtly different from this.
    – pulsar512b
    Commented Aug 27, 2020 at 14:59
  • 1
    "not so much" - huh? Jakovenko was ranked number 5 in the world at his peak rating.
    – Kostya_I
    Commented Jan 3 at 23:07

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.