http://www.chesscafe.com/mueller/mueller.htm

Check out the reader correction of Berg-Kallio, apparently there are exceptions to the Vancura draw...

Well, I did manage to win in 36 moves. People actually started a betting pool to see if I could make it.

After the win, I told about it  to  Jonathan Rowson, with whom I was training back than, expecting praise. Instead he told me how he used to train to give mate within 60 seconds.   

Related to the subject about "technical endgames" people are supposed to know but occur rarely OTB. Two years ago, I played a tournament in which I was 0.5 point behind the leader and had to play him in the last round. A victory would have guaranteed me the overall win of the tournament...

... and it all came down to b+n against 3 pawns; after rounding up the pawns I had about an hour left on the clock to try to mate him. Not only did I never thought that I would have to mate someone with N+B ever, but here I had to do it to win the tournament. 

Suddenly I felt very different about those obscure positions and studied them. If anything, it gave me the confidence to transition into an endgame, where I knew how to defend, or how to play for an advantage, instead of agreeing to a premature draw or even playing something dubious just to avoid that endgame. I was surprised to see how many times it was possible to outplay an otherwise equal opponent in an approximately level endgame.

I mean that white has to keep the pawn on h2, not on h3, because I have this in mind (as black can force one of the following constellations rather easily):
A)
http://www.jinchess.com/chessboard/pos=8/R7/7p/5p2/5P2/2k4P/p5K1/r7%20b%20-%20-

B)
http://www.jinchess.com/chessboard/pos=8/R7/7p/5p2/5P2/2k5/p5KP/r7%20b%20-%20-

In the top position, a transformation to a pawn endgame with 1...Rc1 2.Rxa2 Rc2+ wins, while in the bottom position the same transformation only draws. That's why white has to keep his pawn on h2.  

Is that what you have in mind, John?
http://www.jinchess.com/chessboard/pos=8/6k1/R6p/5pp1/5P2/6P1/p5KP/r7%20w%20-%20...

The position is very close. I think White just barely draws, as long as he keeps his pawn on h2(!). I'll try to post some analysis later.
Let’s just say that for any practical purposes White should have prevented the h6-g5 idea by having played h2-h4 in advance.

I think Black still wins, Pantu. Say White plays Kg3 in the position above, I think actually just an immediate …f4+ is good enough. Whether White plays Kf2 or Kh2 he can only move his rook after that, and Black can follow the general plan of …g4. fxg4 is forced (else …g3+ wins). Then Black can arrange …Kxg4, then he takes the king round to attack e4. Say White has played the tougher Kf2; Black arrives at the queenside and goes 1….Kb3 2 Rb8+ Kc3 3 Ra8 Kd3 4 Ra4 (all forced) and now 4…Rd1 5 Rxa2 Rd2+ 6 Rxd2 Kxd2 (as Ptero suggested above), and wins the pawn ending.

Black could traipse the king round to f1 before he does any of this, perhaps, and then after …f4+ (which now forces Kh2), back to g7 before going …g4 and so on. It’s much easier to win at the end with the king on h2 because Black has e3 and d3 for his king and can zugzwang White almost at once. But I don’t think he needs to.

This does raise another question I’d never considered: would the standard fgh endgame be won if White’s pawns were f4 g3 h2 against f5 g6 h7? Assume Black were allowed to go ….h6…g5 in peace (maybe White can play h4 to stop this, but one thing at a time). White still can’t take on g5 since the f-pawn wins. Would the same plan work? I think not because …h4 is not possible, since White can always prevent you recapturing after gxh4, by Rh8+ or Rh7+. And if you take on f4, then you can’t put him into zugzwang and win the f-pawn, because once you get Ke3 –v- Kg2 with the R on a4, even if it’s White’s move he can go Kh3 and if needed Kh4. You would win if you managed to get him to play h3 before setting this position up. I toyed with trying to force this by …Kh5 threatening ….Kg4, but White can always meet …Kg6 with Ra6+. Now if you go …Kh5 then h3 is fine, because he keeps his rook where it is and plays Kg2-h2 and the Black king can’t get out.

So I think the answer is no. Any endgame wizards want to show me how? Ptero?

IMJohnCox wrote on 02/14/07 at 17:53:01:


Indeed (thanks John, sorry if I wasn't clear)

BTW once the black King gets to the third rank, there are also thematic threats of simplification to pawn endings where black wins by obtaining the opposition, say, K on d3 where black can go Ra1-c1, white replies with Rxa2 and Black plays Rc2+ forcing an exchange

sorry - to get e+g it's the diagram position with pawn g3 and Kg2.  Then black plays fxe4. and after fxe4 Re1 Rxa2 Rxe4 we have the position.

Although thinking about it now instead of Re1 black has Kh6 again.  If he walks the king to g4 before grabbing on e4 his winning chances would be much improved.

But generally I don't think that black can force this position (but it's impossible to be conclusive).  I think in a practical game black should definitely play on and force white to demonstrate the drawing technique in the pawn was on g6.

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.  

I have no modify button but the position i meant of course was 
White Kh2, Ra8. Black K anywhere, Ra1, pawn a2
I mixed up white and black while typing

I believe each chessplayer would benefit much more to study such positions as the vencura and built up a solid endgame techic based on this, than to study openings. Even the position is seldom on the board it will occur often in our caluculation (ok dont play like this, this its a draw etc). 

However I must addmit I spend much more (plenty of) studytime in openings/middlegame and only few in endgames. The reason is simple--> to study openings/middlegame is much more fun. It is the creative part of the chessgame and that part i like most. studying endgame is much more serious. the diffrence between right and wrong is clearer and the consecquences are win or lose. in middlegames there is still lot of play in the position even I am better or worse the outcome of the game is still open. 

Allways when I again failed once more to convert a middlegame advantage in the endgame I think now chessy, stop studying openings and learn this endgames or you well never make progress. But then at home... well I am not a professional I just play chess for fun and so I am again on chesspublishing...

And when I think about the opening lines I analysed but never had on the board....
well maybe with cox`s berlin book I will start to look at end games  

I could probably draw the Vancura position itself, thanks to John Emm's coverage in "Survival Guide to Rook Endings". But I'm not at all sure I would be able to correctly reach it, say with the stronger side's king moving up the board...
 
In practice the endame is a weak point in my play (2150-ish overall), even though I know more "technical" positions than many stronger players I know. I have come to think that studying strategic endings and thinking methods is much more important, since logically they arise more often, and I'm now spending my endgame time with Lars Bo Hansen (Secrets of Chess Endgame Strategy) which looks great so far.

Which is more important in practical play against not-always-perfect opposition: 
Being able to reach technical positions where you are the one with the winning chances most of the time, or being able to play those positions absolutely flawlessly?  

Ok...I just looked up the position with google, and that explains why I didn't recognize it. For that general class of R+p endings, I opted to study only the one or two exceptions where the stronger side could actually win. Then I ran on the assumption that for the most part, common sense would allow the weaker side to draw otherwise, assuming that the weaker side understood the basic principles of R+p.

Something like that, yes. I suppose if I just bring the king to g3 I was concerned he would play ...Ke5...f5; I'd have to take, now my king can't approach his pawns and he can slowly bring the king to the queenside - if the king got to b4 while I was still passive I'd surely lose.

I don't know - I think if I had to panic intuitively again in this position I'd panic intuitively the same way (which did work in the game, at least), but I did mean to have a play with the position and see whether this was in fact either desirable or necessary, and I haven't got round to it.

Vancura isn't the f & h pawn ending, no. Don't know the Euwe book you mention but Keres is good on it too - know what you mean about Fritz, though! That's something else I always mean to do - train some of these classics against Fritz 10 and see how I go. Dreadfully, I imagine.

Well in that case this endgame had already featured some truly embarrassingly bad play by White: the sort of thing that in a just world would have a player’s title revoked. Rc2 was played in time trouble and I’m not at all sure it was a good move. I always meant to analsyse this ending but could never bring myself to revisit it. Woofwoof’s Kf2 looks fairly sensible except for the psychological issues – I’m fairly sure my last move was the hugely inept Kf2-e3…..

IMJohnCox wrote on 02/13/07 at 16:36:56:

It is. 
I was wondering if you would recognise it.

I would rather say that many 'knowledge' endgames are rather common. Perhaps the positions aren't actually happening but that's because one of the players steers away from them or draw/resign happen before they actually appear.

I know very few positions by name. Don't think I've heard of the Vancura.

IMJohnCox wrote on 02/13/07 at 10:54:51:


That matches my experience too, even the R v P one where my opponent got a draw by promoting to a Knight which I hadn't considered when I went into the line.

I've been on the wrong side of R+B vs R and lost it. I looked up the defensive technique afterwards, but might have to wait another 10 years for it to come up again.

I've studied endings quite a lot, and I know I throw away a lot of points through poor ending play but it seems to be done before I get to these theoretical positions. 

In tournaments, often there is one game left with lots of specators and I've seen these theortetical postions come up and seen them misplayed (e.g. a 2000+ rated play who didn't know how to build a bridge) and have thought I'd hate to be in that position, so I think studying them is more for psychological comfort rather than practical use. Hmm, I kind of think that's the same with studying openings.

There are some endgame specialists, but it seems even GMs make a lot of mistakes in endings, and even in their analysis of them, so there's hope for us all.

I have no idea what the Vancura position is. But I must say I have earned numerous half points due to theoretical positions, especially in blitz and rapid.

I was reading Keres' Practical Chess Endings in the bath the other day (as you do) and enjoyed his very classical comment on this and many other R & P -v- R positions that 'a good player must know these by heart'.

Of course few of us would call ourselves good players by Keres' standards, but I just wondered. I would do a poll if I knew how, but just out of interest, how many people here would know (even roughly) what the Vancura position was? Don't post what it is (yet), just whether you'd recognise and/or be able to apply it.

I'll start it off by saying that I would have known the thing properly in about 1982, and that I have done for about the last week, but in between producing the name from a dusty cupboard of my memory if sufficiently prompted (say by a multiple choice quiz) would have been the best I could do.

Is any of this stuff ever actually important? My impression is that few English IMs could manage this either as a matter of knowledge or indeed over the board at a seven-hour time limit (see Kazhgalayev-Hanley, Port Erin 2006).  GMs would vary. I heard Keith Arkell call Vancura 'a trivial draw' but I wouldn't bet on him knowing the name (not that matters, of course). On the other hand I wouldn't back Willie Watson to be able to tell it from a banana, though that's not to say he couldn't work it out if it came up.

But does it matter? I can't think of a single occasion on which I've had any of the classic 'knowledge' endgames - R & P -v- R, R f&h -v- R, R and a pawn -v- R with fgh each, R & B -v- R, those R & P's -v- B & P's fortresses, Q -v- R, Q -v- R and pawns fortresses, R -v- 3 Ps mutual zugzwangs, and so on. I suppose I've had the odd Lucena, and once a R v Pawns endgame with a thematic trick (which I didn't know but stumbled into). Maybe I've just been lucky and I'm in for a horrible retribution in due course.

What are other people's experiences about this?