Quote:


Do you mean with Black still having the a-pawn? In that case I think it is won for Black:
a) The pawn marches to a2 and White's Rook cannot leave the a-file,
b) the e-pawn is used as a distraction for the White King, that will have to leave the g2-h2 area,
c) the Black Rook goes to h1 and ..bingo!

The pawn got to g5 on move 24 - gaining space like that is usual I think in these endings.  I don't think we can say that threw away Jacob's winning chances as you were still a pawn up at that stage   - although I think it was going to drop anyway.

If my understanding of Dvoretsky is correct you got the spirit right (sac a pawn to convert a messy inferior position to a more exact drawn one) but I think Rc7 is inferior because of Rb4.

[Event "14th Monarch Assurance"]
[Site "Port Erin IOM"]
[Date "2005.10.02"]
[Round "9"]
[White "Cox, JJ"]
[Black "Aagaard, J"]
[Result "1/2-1/2"]
[WhiteElo "2385"]
[BlackElo "2442"]
[ECO "E25"]
[EventDate "2005.09.24"]

1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.f3 d5 5.a3 Bxc3+ 6.bxc3 c5 7.cxd5 Nxd5 8.dxc5
Qa5 9.e4 Ne7 10.Be3 O-O 11.Qb3 Qc7 12.Nh3 Nd7 13.Nf2 Nxc5 14.Qc4 b6 15.
Bxc5 bxc5 16.Nd3 Nc6 17.Qxc5 Ba6 18.Nb4 Bxf1 19.Nxc6 Rfc8 20.Rxf1 Qxc6 21.
Qxc6 Rxc6 22.Rc1 Rb8 23.Kf2 Rb2+ 24.Kg1 g5 25.Rfd1 Ra2 26.Ra1 Rxa1 27.Rxa1
Rxc3 28.Ra2 a5 29.a4 h6 30.Kf2 Kg7 31.Kg3 Kg6 32.h4 Rc4 33.hxg5 hxg5 34.
Kf2 Kf6 35.Ke3 Rb4 36.Rc2 Rxa4 37.Rc7 Ra1 38.Kf2 a4 39.Ra7 a3 40.Kg3 Ra2 
41.Ra5 Kg6 42.Kh3 f6 43.g3 Kf7 44.Ra7+ Ke8 45.Kg4 Ra1 46.e5 fxe5 1/2-1/2

@IMJohnCox: That's what's in the database.  Aagaard then played Ra1 allowing Ra7 when I'd expect it to be drawn.

@Ptero: I think perhaps you are quoting the wrong part?  I say that fgh is safe and you disagree by claiming efg is much more dangerous?

I'm not too sure about your idea with the pawn on g6: it's certainly true that with (say) black e+f vs white f+g he has more chances but I don't think he can obtain this unless white does something silly - at the least white can just let black take on e4 and recapture with his f-pawn.

For instance: take the position after my suggested 37.Rc5, and put the black king on f6 and the g pawn on g6.  How is black to proceed?  

Also if I remember my theory correctly the Vancura position where the attacker has a+g pawns is drawn too (a+f is a win).  So if from my suggested position we have 37...e5 38.g3 Ke6 (intending ...f5) 39.f4 there follows either 39...f6 40.fxe5 fxe5 which should be drawn or 39...exf4 40.gxf4 and white sits tight in this position (40...Kd6 41.Rd5+ Kc6 42.f5 should be ok?).  If black forces a passed g-pawn white can just throw away the e-pawn, stick his king on g2 and await the draw.

If black doesn't play ...e5 then white also has the option of playing e5 in response to f5.

I don't think you can give an definitive assessment of these types of positions as there are too many ideas and counter ideas, but I "believe" that if white reacts correctly it should be drawn.

Quote:

I don't agree. The constellation with efg, rather then fgh is more dangerous to the defender. In the afore mentioned game, If the black's pawn was on g6 rather then on g5 (when black has a passed a-pawn and both sides have efg pawns) he would have had serious winning chances based on f7-f5, as exf5 could be answered by gxf5 with the possibility of creating a second passed pawn. As the position stands with the pawn on g5 it should most probably be drawn.
I agree that DEM is an excellent source for studying these positions. Another excellent source is John Emm's "survival guide".

@IMJohnCox:  Dvoretsky and Yusopov cover this a-pawn rook ending in great detail in Technique for the tournament player (I think the ChessCafe articles are based on this).

They start off with the drawn position with pa7, then Tarrasch pa6 followed by the actual drawing method.  Then it is considered in various practical endings forcing this.  This includes a Karpov-Yuspov ending from a linares (http://www.chessgames.com/perl/chessgame?gid=1068855) and a few others where black has a pawn somewhere else (so level material) but he sacs this pawn to force this position.  Basically it's a treatise on how to utilise the theoretical knowledge of this ending in practical play to convert an inferior (although drawn) complex unknown position into a position with exact knowledge.

My understanding is that Dvoretsky's Endgame manual is supposed to hold all the positions for which you should have this exact knowledge and you should then apply it in the same way.  I was frustrated by Technique for the Tournament player precisely because it blithely states that such exact knowledge is essential for any strong player but doesn't tell you what they are!  D.E.M. solves that problem.

Re you comment about Miles not recognising it: they quote an article in wrote in an American magazine by him regarding this ending with a couple of extra pawns elsewhere claiming that he'd shown convincingly it was a win despite what theory said.  Dvoretsky quotes this and of course shows it's a draw.  Several other famous GMs get their wrists slapped for this to.  I can't be more precise as this book is in the attic of my brothers house a couple of hundred miles away.

Specifically with regards Cox-Aagaard I would still go for 36.Rc2! but after 36...Rxa4 everything I've learned about these positions suggests 37.Rc5! and white has good chances to hold.  With fgh instead of efg I don't think there are any particular problems provided the white king is either on the 2nd rank or covered from the checks (otherwise Ra1 with pa2 isn't drawn!).  Shifting the pawns might give practical chances for black though but I'd bet on white holding.

Specifically I think 37.Rc5! is better than 37.Rc7? because of 37...Rb4! when 38.Ra7 a4 is an ending where the black rook can influence matters on the kingside.  It's possibly still drawn but black should have better practical chances?

P.S. picked up Dvoretsky's Endgame Manual and the section on this ending is in there - among the tragicomedies is Ivanchuk-Lautier, Horgen 1995 where both blundered several times.  Also in there is Emms-Riemersma, Gausdal 1993 which white lost from a drawn position - obviously this was before he wrote Survival Guide to Rook endgames...

Quote:


John, this is exactly what I was thinking when first looking at this position. If Black keeps his pawn at a3 for a while, he may be able to return on time to catch White's K-side pawns, my logic is as follows:
a) try to arrange the Black K-side pawns in a way that not all of them are going to drop to the White Rook & march with the King to support the a3-pawn,
b) then White loses his Rook for the a-pawn,
c) the Black King & Rook rush back to catch the White pawns before promotion,
The main problem I see with the above, is in case Black loses all of his 3 pawns! I haven't played this with a board, but it doesn't feel to me like a certain draw. What do you think?

Another more adventurous plan for Black could be to try to make the e-pawn a passer. Then even with the pawn on a2, Black can win as the e-pawn will distract the White King from his g2-h2 refuse..

Just a thought, chk

P.S.: I have only seen the Vancura position in Wikipedia & in a test (in chessnia I think), but I never found the time to explore it. However, I do believe in slowly building endgame knowledge by learning some standard positions that you think may help you in decision-making in the endgame. For instance, you can read Sheresevsky's Endgame Strategy 5 times and still not be able to understand practical Rook endings without learning some standard technique/positions, because so many of them are counterintuitive as you have already noted.. Moreover, pawn endings and Queen endings are really the cornerstone to all other endings, so you cannot ignore them either. And lastly minor pieces! If you are an expert on those, then not only you win points in the endgame, but you also improve your overall middlegame skills..

So, overall, no time is wasted in gathering chess knowledge, you just have to know where to draw the line. On principle, one should try not to overdo it with openings and same applies to endgames and everything else. But in the end, this is all a matter of taste I suppose.. Does it affect your game? Well, everything does..

I learned the Vancura position rather recently when I made an effort to shore-up my scant endgame knowledge.  I did so because most players I encounter (OTB - internet is a different matter) are ex-soviets, and I always feel profound inferiority in basic chess knowledge when playing them.