King & Queen v King & Rook ending

[Ian Kingston]

In Nunn's Secrets of Pawnless Endings (1994), p. 67 he gives this position as the longest win:
[FEN "K3r3/8/5k2/Q7/8/8/8/8 w - - 0 0"] [PlyCount "61"] 1. Kb7 Re7+ 2. Kb6 Re6+ 3. Kc7 Re7+ 4. Kd8 Re4 5. Qc5 Re5 6. Qc4 Kf5 7. Kd7 Re4 8. Qf7+ Ke5 9. Qh5+ Kf4 10. Kd6 Rd4+ 11. Kc5 Rd3 12. Kc4 Rg3 13. Kd4 Rg5 14. Qh7 Rg4 15. Qd3 Rh4 16. Qf1+ Kg4 17. Qf7 Rh2 18. Ke3 Rh3+ 19. Ke4 Rg3 20. Qh7 Kg5 21. Qh2 Kg4 22. Qh6 Rg2 23. Qg6+ Kh3 24. Qh5+ Kg3 25. Ke3 Rg1 26. Qe5+ Kg2 27. Ke2 Kh3 28. Kf2 Rg2+ 29. Kf3 Rg4 30. Qf5 Kh2 31. Qxg4
Nunn gives a different line (according to my computer, White has no fewer than 11 equally long wins to choose from at move 5), but it still takes at most 31 moves to win the rook.

[Kevin O'Rourke]

i once wondered if 2 knights can hold their own against a Queen because I thought maybe they could defend each other and keep the king nearby.

turns out you can't hold it. Sooner or later you have to move a knight or be checkmated. One knight falls and then the other.

[IM Jack Rudd]

i once wondered if 2 knights can hold their own against a Queen because I thought maybe they could defend each other and keep the king nearby.

turns out you can't hold it. Sooner or later you have to move a knight or be checkmated. One knight falls and then the other.

I just checked it on Nalimov; it's a draw.

[Roger de Coverly]

I just checked it on Nalimov; it's a draw.

There are some winning positions. This one, generated by playing around with Shredder

5Q2/8/8/4n3/3k4/5n2/5K2/8 w - - 0 1

[JustinHorton]

There are some fairly recent instances of super-GMs failing to put this one away, of course.

Like for instance

[Kevin O'Rourke]

i once wondered if 2 knights can hold their own against a Queen because I thought maybe they could defend each other and keep the king nearby.

turns out you can't hold it. Sooner or later you have to move a knight or be checkmated. One knight falls and then the other.

I just checked it on Nalimov; it's a draw.

so the 2 knighs can hold the draw? nice one. must remember that

[Paul Dargan]

Q v R:

I remember a chessbase article that spoke about a guy that had done a lot of work on the ending, including a database of key positions. Seem to recall 'rosettes' and other patterns being described.

The guy was called Derek Grimell ... His site seems to have been taken down now though.

The previously metioned Gelfand-Svidler game and commentary.
http://en.chessbase.com/post/can-you-play-this-endgame-



Paul
[Edit: worst spelling and formatting issues fixed]

[Stewart Reuben]

There are available databases for all endgames upto and including seven pieces.
The most interesting from my viewpoint as a law maker is K+Q+N v k+r+b + n.
In this position, with optimum player on both sides, White can win on move 545, but makes the first capture on move 512.
White Kd3. Qh1. Nh2. Black Kf4, Ra7, Nh7, Bd1. Black to play, http://chessok.com/?page_id=27966
Of course it falls foul of the 50 Move Rule. The reason I did this research was concerning the nature of that rule. Both players could perfectly reasonably play on and on in an effort to win and neither claim on the 50 move rule. The Laws since 2014 means that the arbiter steps in after 75 moves to declare the game drawn. The reason for this is that, with an increment of 30 seconds, games could go on and on.
We chose 75 in ignorance of the recent database discoveries. Probably aesthetically it should be a higher number, perhaps 100 or 200.

Some years ago there were challenges of Q+R human against K+R computer. GM Walter Browne failed to win the first game. I think it was blitz.

[Clive Blackburn]

Some years ago there were challenges of Q+R human against K+R computer. GM Walter Browne failed to win the first game. I think it was blitz.

I realise that it was probably just a typo but what was the material balance for these challenges please Stewart?

[Clive Blackburn]

Oh ok I Googled it, it was K+Q v K+R.

Here is the game that Browne failed to win:-

http://www.chessgames.com/perl/chessgame?gid=1480950

[Brian Towers]

The Laws since 2014 means that the arbiter steps in after 75 moves to declare the game drawn. The reason for this is that, with an increment of 30 seconds, games could go on and on.
We chose 75 in ignorance of the recent database discoveries. Probably aesthetically it should be a higher number, perhaps 100 or 200.

You cruel, heartless beast! Don't you think arbiters have a home to go to? With a warm bed waiting?

[Stewart Reuben]

That is why I didn't suggest 512.

[NickFaulks]

We chose 75 in ignorance of the recent database discoveries. Probably aesthetically it should be a higher number, perhaps 100 or 200.

I didn't think that was the point at all. I thought 75 was chosen because it is safely well past 50.

[Jon Mahony]

The one time I had it, I couldn’t do it and after messing around for ages, I wound up just taking the Rook with my queen in frustration making it a K vs K draw. I did some studying up on it afterwards, but this was a couple of years ago, not sure I’d be able to win it again now.

[Andrew Sainsbury]

Hi All

I have written up the analysis I have done on this ending. Hopefully this is useful to some of you.

Part 1 - Winning from the Philidor Position
http://www.thequietmove.com/queen-v-roo ... -philidor/

Part 2 - Getting to Philidor Position
http://www.thequietmove.com/queen-v-roo ... -philidor/

There's really no short-cut for this, as the ending is complex with best play from the defender. In practice, you may find a nice fork along the way!

Would appreciate any feedback.
Best wishes,
Andrew