NH Chess Tournament 2010

[Matt Mackenzie]

It was against Gashimov in a Spanish League game (if you have the latest Informator, it's in there)

[Richard Bates]

Secret to experiencing lots of quick losses - continue to play the sharp openings of your youth, whilst not putting in the volume of preparation work that you used to do.

[Jonathan Rogers]

John Nunn wrote an amusing passage in his second games collection about this, adding that one should really play Black too.

Anyway, after everything I said yesterday, "naturally" Nielsen went on to win a nice game against Giri today!

[Richard Bates]

John Nunn wrote an amusing passage in his second games collection about this, adding that one should really play Black too.

My experience suggests that this isn't remotely necessary

However, I think after this tournament David may be appreciating the attraction of losing quickly. Another endgame for juniors (and, er, the rest of us) today.

[Jonathan Rogers]

Yes, indeed. Can anyone explain to me how one knows whether one is winning an endgame with queen against rook and pawn, assuming that the rook, pawn and king are all within close proximity of each other? Presumably whether or not there is a fortress depends on which rank and file the pawn is, but exactly what are the principles?

[IM Jack Rudd]

A rook's pawn is the worst, because then you can't do the "shuffling rook" defence. (King guards the pawn, rook moves back and forth between the two squares where the pawn protects it.)

[Jonathan Rogers]

Why not? And why did this defence still fail today with the pawn on e6?

[Richard Bates]

Why not?

Er, because a rook's pawn only controls one square? Also rook's pawns don't provide cover for the K to hide behind.

And why did this defence still fail today with the pawn on e6?

Because of what happened. The extra rank allows the queen&K to use zugszwang to drive the black K in front of the pawn and allow the white K to cross the fifth rank and approach the pawn, after which it can't be held. The fortress positions result from the black pawn being on the seventh rank. Here is a recent example: http://www.michaeladamschess.co.uk/2010 ... ortress-2/

Apparently Philidor analysed all this over 2 centuries ago!

[Jonathan Rogers]

oh, does that mean that the pawn absolutely has to be on the seventh rank, then (except perhaps a7 or h7)?

[Richard Bates]

oh, does that mean that the pawn absolutely has to be on the seventh rank, then (except perhaps a7 or h7)?

Well i suspect there may be other specific positions where the white K is completely cut off eg. Black pawn on f6, R on g5, white K on h4, but in general that would be my guess. But it is only a guess!

EDIT: actually this seems to be a fairly simple win as well, using same technique - force the Black king to f5 and play Qe8.

[Christopher Kreuzer]

oh, does that mean that the pawn absolutely has to be on the seventh rank, then (except perhaps a7 or h7)?

Well i suspect there may be other specific positions where the white K is completely cut off eg. Black pawn on f6, R on g5, white K on h4, but in general that would be my guess. But it is only a guess!

I suppose you could ask Philidor instead of guessing...

As these are 5-piece positions, can't you use tablebases as well?

My guess (yeah, I don't have tablebases or Philidor either) would be that the more space there is around the R+P+K structure, the more chance there is for the K+Q to approach and disrupt it in some way. I suspect queening lines also help the draw. The position where the defending side has the pawn on b2 and K on a1 and the rook shuffling between a3 and c3, would surely be a draw as the attacking king couldn't approach, and the queen can't check from behind (as it would be able to do if the pawn was further up the board).

So the geometry seems to be for the rook to shuffle between the two defensive points of the pawn (as Richard said) and keep the attacking King out, and for the defending King to use the edges of the board to restrict the lines by which the queen can attack. Anything out in the middle of the board seems doomed.

Though if someone could point to a current book where this is covered, that would help!

[Jonathan Rogers]

Agreed, this is not well covered in the books. I would have asked Philidor had he been a bit more lively, but as it was we didn't even get an answer from Keith ...

In fact I have only seen anything remotely relevant to this covered in a book; it was in a book of lessons (not specifically endgame lessons) by Kasparov, though quite possibly written partly by a ghost writer, in the late 1980s. This was a short thematic book with most of the diagrams and continuations aimed at players around 150, or so it seemed to me when I flicked though it at my local bookshop (of course one would never browse instead of buying at Chess & Bridge). But one of the diagrams had queen against rook and two connected pawns - pawns on c5 and d4, I think. Even this was also said to be winning. It seemed rather out of place in the book though!

One day soon I will set up queen v rook and e6 pawn on a board and feel my way towards winning it, Svidler style.

[Christopher Kreuzer]

Here is a recent example: http://www.michaeladamschess.co.uk/2010 ... ortress-2/

Apparently Philidor analysed all this over 2 centuries ago!

Thanks for the link to that game. I hadn't realised Michael Adams has a website. From the comments to that game, it seems the website is new and it looks very nice! Do many grandmasters now have their own websites like this?

I went and Googled the Philidor reference, and a search for the (separate) terms "philidor" "queen" "rook" "ending" led me to the following:

http://en.wikipedia.org/wiki/Philidor_position

The "Queen versus rook" section there doesn't include a pawn, but there is more in this article:

http://en.wikipedia.org/wiki/Chess_endgame_literature

[A fascinating aggregation of endgame sources, if you are prepared to use it as a starting point and take what is said there with a pinch of salt until you can lay your hands on the actual sources used in the article]

"Philidor's book [1749] contained much more endgame analysis than earlier books. The first edition analyzed the rook and bishop versus rook endgame. Later editions covered [...] queen versus rook and pawn".

So it is a "later edition" of the 1749 book by Philidor. Toddling over to the Philidor article:

http://en.wikipedia.org/wiki/Fran%C3%A7 ... n_Philidor

"In 1749, Philidor published his famous book Analyse du jeu des Échecs. He printed a second edition in 1777, and a third edition in 1790. The book was such an advance in chess knowledge that by 1871, it had gone through about 70 editions, and had been translated into English, German, Russian and Italian."

So if anyone sees a copy of this book (Analyse du jeu des Échecs), have a look at what edition it is and see whether it includes the rook and pawn vs queen analysis!

A bit more Googling threw up two more webpages.

This one is not that helpful (not very specific):

http://www.westmichiganchess.com/author ... aspx?ID=38

"try clustering your king, rook and pawn close together setting up a defensive fortress to draw by the 50 move rule (and don't move your pawn)."

This one is more helpful:

http://gameknot.com/room-fm.pl?home=2&th=121

Three positions there. Hopefully no-one reading that will ever now go wrong on either side of this endgame...

Apparently the book to get hold of is this one: "Lehr und Handbuch der Endspiele" - Band 3, 1969, by Cheron. Should be easier to get hold of than editions of Philidor's work (though I hope modern reprints of Philidor are available).

"Cheron mentions that his research on the K+Q vs K+R+p endgame took him more than a year of hard analysis work."

[Andy Burnett]

There is plenty of coverage of these types of positions in my copy of Batsford Chess Endings (Speelman/Tisdall and Wade). One of the positions given dates from a 15th century Spanish manuscript!

The winning/drawing positions are not quite as clear as they might appear to be at first - e.g. there are 3 similar positions with rook's pawns where if the pawn is on the 2nd rank it loses, on the 3rd rank it draws and on the 4th it loses again!
Quite interesting to play through some of these but perhaps of little practical value (I don't recall reaching such a position or needing this knowledge in 25 years of playing).

[Maxim Devereaux]

Practical Chess Endings by Keres has a good section on Q v R+P. The basic conclusions are: if the pawn is on the 7th, it's generally a draw. Otherwise, with a central pawn, it must be on its original square in order to draw (usually with the "shuffling rook defence as suggested by Jack), a bishops pawn provides marginally better defensive chances, a knights pawn significantly better (due to shelter / lack of space for zugzwang manouevres with the queen), but rooks pawns are quite often, though not always, lost.

For those who don't have the Chessbase Endgame tablebases, the Nalimov tablebases up to 6 piece endings can be accessed online here