English Chess Forum

K &Q v K & R

Is this confirmed as a win for the Queen against all defences possible? I thought that since the invention of strong chess computers it was shown that the queen can win everytime. Pre engines though was it possible to get escape with a draw?

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

Is it a confirmed win against all defences?

If I put it into a computer strongest level it would beat itself?

https://en.wikipedia.org/wiki/Pawnless_ ... ersus_rook

Queen wins in all positions without obvious draws/wins (e.g. stalemate, captures, perpetual check).

Tablebases can be used in such endings if you want to explore the best moves.

I use the online Shredder site:

http://www.shredderchess.com/online-che ... abase.html

Back in the days when adjudication was common, does anyone remember how these were treated?

Kevin O'Rourke wrote:K &Q v K & R

Is this confirmed as a win for the Queen against all defences possible? I thought that since the invention of strong chess computers it was shown that the queen can win everytime. Pre engines though was it possible to get escape with a draw?

There are a handful of drawn positions (excluding those where the rook can capture the queen immediately). Perpetual check or stalemate are the mechanisms. This example shows both:
[FEN "5k2/5r2/4Q3/8/8/8/8/7K b - - 0 1"] [PlyCount "11"] 1... Rh7+ 2. Kg2 Rg7+ 3. Kf3 Rf7+ 4. Kg4 Rg7+ 5. Kf5 Rf7+ 6. Ke5 (6. Kg6 Rg7+ 7. Kh6 (7. Kf6 Rg6+ 8. Kxg6 {stalemate}) 7... Rh7+ 8. Kxh7 {stalemate}) 6... Re7
The endgame was solved by databases a long time ago.

Kevin O'Rourke wrote:K &Q v K & R

Is this confirmed as a win for the Queen against all defences possible? I thought that since the invention of strong chess computers it was shown that the queen can win everytime. Pre engines though was it possible to get escape with a draw?

Pre tablebase, it was thought that it was always a win with correct play. Post tablebase, it became apparent that there were some previously unconsidered defensive ideas which could prolong the game. If the defender can survive 50 moves, it's a draw.

Shredder is as good a way of any of checking various positions. It's usually mate in not much more than 30, so it's only if the side with the queen spends 10 or 20 moves getting nowhere that the draw becomes a serious possibility.

http://www.shredderchess.com/online-che ... abase.html

Very occasionally the rook will win due to a skewer. Imagine that.

NickFaulks wrote:Back in the days when adjudication was common, does anyone remember how these were treated?

It would be marginally surprising if it ever arose. Adjudication may have been common, but endings weren't, particularly ones that it would be unlikely ever to reach between moves 30 and 40. Wouldn't many beginner's books have covered Queen v Rook in one of their endings chapters? Golombek's "The Game of Chess" does, and throws in the drawing position (above) for good measure.

You could list all the possible (legal) positions and give them as either a draw, win for the rook or win for the queen. I am sure that such an analysis will be available somewhere. You could break it down further into immediate captures winning, immediate tactics (e.g. skewer or forks) winning, and ditto for immediate draws. It is useful to know the various rook 'defences' to prolong the game if the 50-move-rule draw is a possibility. Ditto it is useful to know the right ways to win.

Does anyone know how common this endgame is in OTB play (according to the databases)?

Christopher Kreuzer wrote: Does anyone know how common this endgame is in OTB play (according to the databases)?

I found around 2200 from (edit) five million(/edit) . But that needed filtering, since many were draws arising from R v queening pawn. Searching purely for decisive results eliminated around 800.

There's a few where the Queen failed to win. The Lahno game being one.

http://www.chessgames.com/perl/chessgame?gid=1658927

There used to be an internet article on this ending (in the days when all chess discussion was via rec.games.chess) which stated that computer analysis had discovered that the "floating rook defence" was the most difficult to overcome in practical play. I think I have seen the same statement in one of John Nunn's ending books (Pawnless Endings?)

A brief play-around with Nalimov leaves me unable to find a position where the forced win takes longer than 35 moves.

Yes it's definitely a win, but the key move(s) to break down the floating rook defence are not all intuitive (in my view). I tried to learn this ending years ago by playing against Fritz, but I couldn't do it. In comparison mating with B+N is a piece of cake (and I actually got it once in a tournament game). Of course it is possible in principle to memorise the move sequence from the basic position(s) and I tried that (with success).

Er, WITHOUT success, I meant.

(But in practice the humans I play don't know the floating rook defence so it's all a bit academic.)