PDA

View Full Version : Endgame looks winning, but how do you win it?



Zwischenzug
17-06-2007, 07:58 PM
So in the position below am I winning or is it leaning towards a draw? You be the judge:

8/3k4/r2P4/1p1PK3/8/P3R3/8/8 w - - 53 53
White to move.





P.S. &?$%#!?!! Curses! I drew this game!

eclectic
17-06-2007, 08:29 PM
I would play Rh3 still protecting the a pawn but also stopping capture of d6 pawn by rook (Rh7+ would then win rook) the white king and d5 operate as a phalanx to drive the white king to the back of the board with this check

Kevin Bonham
17-06-2007, 09:57 PM
This looks more difficult than it looks, if that makes any sense. ;)

Trent Parker
17-06-2007, 11:32 PM
ok I analysed this posi with Zwischenzug afterwards and came up with:
1. Rf3

Now obviously
1....Rxd6 loses to 2.Rf7+ Ke8 3.Kxd6 Kxf7 4.Kc7
lots of variations though,......

Igor_Goldenberg
18-06-2007, 12:15 AM
At the risk of making a fool of myself:

1. Rh3 Ra4 (anything else seems to lose on the spot).
2. Rh7+ Kd8
3. d6 White is threatening Ke6, then Ra6+ is met with d6 (no stalemate tactics because of b5 pawn!), Re4 (or Re3 if black takes pawn on a3) is answered by Kd6. Two sensible defences for the black:

a) 3...Kc7 4.d8Q+ Kxd8 5.Ke6. Looks like win for white because of a3 nad b5 pawn
b) 3...Ra6 4.Rg7 (important waiting move to take the threat from a3 pawn.) If black plays Rb6 or g6, then Kd4-c5, winning b5 pawn.
alternatively, black can play 4...Kc7, but then white has:
5. d6+! Rxd6 (5...Kd8 6.Ke6) 6.d8Q+ Kxd8 7.Kxd6

Overall it seems whinning for white, but an analysis with the board and/or computer/tablebase might find something I missed.

Igor_Goldenberg
18-06-2007, 12:26 AM
A little refinement (from iron friend):
After 3...Ra6 white can play 4.a4!! bxa4 5.d6 winning (rook cannot check white king from the rear).
4. Rg7 might not be sufficient after 4...Rg6 5.Kd4 Rg3 6.Kc5 Rxa3 7.Kc6 Ra6+

Endgame after 3...Kc7 4.d8Q+ Kxd8 5.Ke6 indeed must be won for white, but very acurate play is required.

Bill Gletsos
18-06-2007, 12:35 AM
I assume you mean 3. d7 as the pawn is already on d6.

Bill Gletsos
18-06-2007, 12:43 AM
Endgame after 3...Kc7 4.d8Q+ Kxd8 5.Ke6 indeed must be won for white, but very acurate play is required.Is that really that difficult.
What can black do?

Phil Bourke
18-06-2007, 01:28 AM
Hey how about that guys :) I hit on the idea of pushing d pawn through to Queen and then Ke6 in the post game analysis, we (zwisch, Trent, and myself) were sure that it was winning, just difficult.

Igor_Goldenberg
18-06-2007, 09:35 AM
I assume you mean 3. d7 as the pawn is already on d6.Yes, typo.

Igor_Goldenberg
18-06-2007, 09:38 AM
Is that really that difficult.
What can black do?1. Rh3 Ra4
2. Rh7+ Kd8
3. d6 Kc7
4.d8Q+ Kxd8
5.Ke6 Rd4
6.Kd6 Ke8
7.Rh8 Kf7

Play against the computer (show analysis is off:) ) and see how many times and how quickly you win.

Kevin Bonham
18-06-2007, 11:55 PM
I'm convinced.

Bill Gletsos
19-06-2007, 12:09 AM
I'm convinced.Convinced that it is easy to win after 5. Ke6 or that it is difficult to win.

Kevin Bonham
19-06-2007, 12:26 AM
Just convinced that it is won.

Aaron Guthrie
19-06-2007, 12:28 AM
If black magically, or even just by normal chess swapping off, removes the b and a pawns at the end of Igor's line it's a theoretical draw.

edit-Ok I've been mucking about with 6-man table bases http://chess.jaet.org/endings/ for a while now and after Rb8-Ra4-Rxb5 its a tablebase win.

Capablanca-Fan
19-06-2007, 03:47 AM
If black magically, or even just by normal chess swapping off, removes the b and a pawns at the end of Igor's line it's a theoretical draw.

edit-Ok I've been mucking about with 6-man table bases http://chess.jaet.org/endings/ for a while now and after Rb8-Ra4-Rxb5 its a tablebase win.

Yeah, the draw would be WR on d8, where it is likely to be if the position had started as R+P v R, and Black's K going to the "long side". The rule is for the K to go to the short side to leave the long side for the R, but Black can get away with it against a centre P. E.g. Kc7 Ra7+, Kb6 Ke7!.

Igor_Goldenberg
19-06-2007, 09:19 AM
If black magically, or even just by normal chess swapping off, removes the b and a pawns at the end of Igor's line it's a theoretical draw.

edit-Ok I've been mucking about with 6-man table bases http://chess.jaet.org/endings/ for a while now and after Rb8-Ra4-Rxb5 its a tablebase win.

During the tournament game you don't have access to tablebase (at least not supposed to). Some position maybe won with tablebase, but difficult to do during a tournament game. Some are a theoretical draw, but next to impossible to draw OTB

Phil Bourke
19-06-2007, 09:52 AM
During the tournament game you don't have access to tablebase (at least not supposed to). Some position maybe won with tablebase, but difficult to do during a tournament game. Some are a theoretical draw, but next to impossible to draw OTB
Indeed it does Igor. That little thing called a clock has a strange effect on what knowledge is readily available :) Zwisch's position is an ideal example, he settled for a draw even though he thought that there was a win there, because he couldn't work out how it was won. It took him, Trent, and I nearly twenty minutes going over the possibilities before we hit on the winning plan. So grabbing us for some quick 'in house' analysis during the game would have only cost him the win on time ;) Perhaps a friendly FM could be at these tournaments ready and armed with these winning lines when we need them most :)

Garvinator
19-06-2007, 11:02 AM
Perhaps a friendly FM could be at these tournaments ready and armed with these winning lines when we need them most :)But Phil, what kind of FM :whistle:

eclectic
19-06-2007, 11:05 AM
But Phil, what kind of FM :whistle:

the kind that ought to be stripped of their title and ratings if caught offering assistance during a game? :whistle: ;)

Aaron Guthrie
19-06-2007, 01:56 PM
During the tournament game you don't have access to tablebase (at least not supposed to). Some position maybe won with tablebase, but difficult to do during a tournament game. Some are a theoretical draw, but next to impossible to draw OTBThe first bit of the post was meant to indicate a practical way to fight for a draw. The tablebase edit was just to confirm that it really is won.

And yes I agree it can be very hard to win (or draw a draw) otb in these endings. Even if you know the technique you can start second guessing yourself after being at the board for a few hours. I had one of these 2 pawn vs no pawn endings where the simplest was to transpose to a winning 1 pawn ending, which I did but only after some hesitation.

MichaelBaron
20-06-2007, 04:27 PM
I think any move is winning for white here...
even 1. Rb3 will do the job!

Igor_Goldenberg
20-06-2007, 04:53 PM
I think any move is winning for white here...
even 1. Rb3 will do the job!
1...Rxd6 is actually a draw

MichaelBaron
20-06-2007, 07:51 PM
1...Rxd6 is actually a draw


two pawns down? wow. I should spend more time with Telebase i guess and learn endgames better :(. I could have easily miss the win with Rb3 in this position

Kevin Bonham
20-06-2007, 08:04 PM
It's a real trap for (not just) beginners - the fact that there are some rook endings even two pawns down that are draws, and others that are very hard to win. Club players often know about possible draws like bishops of opposite colours two pawns down (I even drew one down three once) or KBP v K where the pawn is a rook pawn and the bishop is the wrong colour, but many are not aware of the number of drawn KRPP v KR endings.

I'd recommend any club player who hasn't already done so to check out the relevant sections in "Basic Chess Endings", "Batsford Chess Endings" or similar.

Igor_Goldenberg
21-06-2007, 10:08 AM
Philidor position is a draw. Adding a3 pawn might not improve winning chances that much. I suspected it might be a draw and decided to check with Tablebase.

Phil Bourke
22-06-2007, 04:05 PM
To continue this thread, I give you the following position from my Round 5 game against John Papantoniou.
8/6pp/2pk4/pp1p4/3P1PPP/4K3/PP6/8 b - h3 0 31
It is my move as Black, and I missed the opportunity to draw this.

MichaelBaron
22-06-2007, 06:21 PM
To continue this thread, I give you the following position from my Round 5 game against John Papantoniou.
8/6pp/2pk4/pp1p4/3P1PPP/4K3/PP6/8 b - h3 0 31
It is my move as Black, and I missed the opportunity to draw this.

was it h5?

Kevin Bonham
22-06-2007, 07:33 PM
I'm not convinced that that position is drawn. I suspect it's actually a difficult win for white.

Any drawing lines?

Capablanca-Fan
22-06-2007, 07:53 PM
was it h5?
Yeah, three pawns abreast are often vulnerable to a strike against the middle one. ...c5 seems bad after dcx5 Kxc5, f5 leading to a typical outside passed pawn win. Is this how the game proceded?

Kevin Bonham
22-06-2007, 08:57 PM
1...h5 2.gxh5 Ke6 3.Kf2!! and black is lost; Black can not prevent White's king reaching g4.

If 2...Ke7 3.f5! and white again wins because once a position with Kf4 vs Kf6 is acheived White runs black out of moves with the maneuver h6! forcing ...gxh6 then White plays h5 - hence White gets Ke5 winning.

Igor_Goldenberg
22-06-2007, 09:54 PM
My first impresion as that black is lost, unless 1...h5 changes the course of the game. More detailed (and I have to admit - computer assisted) analysis revealed that it was wrong on both counts.

After 1...g6 the game is drawn. Indeed 2.f5 gxf5 3.gxf5 Ke7 leads to a position where white cannot break through.

1...h5 is actually loses, as Kevin pointed above.

Second critical look at the positon revealed what computer missed: 1...g6 also loses after 2.f5 gxf5 3.g5! Having not beleived computer could not spot such a simple move I decided to give it more time. After some churning of the processor the iron wizard found it.

The verdict - position is lost for black

Capablanca-Fan
22-06-2007, 10:02 PM
Well done!

Kevin Bonham
22-06-2007, 10:09 PM
Second critical look at the positon revealed what computer missed: 1...g6 also loses after 2.f5 gxf5 3.g5! Having not beleived computer could not spot such a simple move I decided to give it more time. After some churning of the processor the iron wizard found it.

Yes - I also found this 3.g5! idea hence my initial opinion that the position is a win for white.

Phil Bourke
22-06-2007, 10:42 PM
Yes - I also found this 3.g5! idea hence my initial opinion that the position is a win for white.
All correct, the position is a win for White. After the game, John P suggested ...h5 as being worrying to him because of reasons first discussed by you, but without any solid analysis I thought that ...h5 was a loss also.
After doing the usual Fritz analysis and nothing being revealed there :) I went down that route of switching on Infinitive Analysis and letting the machine chew away.
There is a drawing line for Black after ...h5 2 gh5 Ke6 but in this line White can win by 2 Kf3!
In the game, I gambled on holding by ...Ke6 and soon after resigned :) I threw it up as a tempter for the machine addicts to grab and suggest solidly that Black can draw, as it is one position that took my 'puter an hour or more to work out.
Now for another position from the game that I may actually have held! Interested in your opinions here, as my knowledge and machine assistance hasn't been insightful at all :)
8/1p4pp/2p1k3/p2p4/3P1P2/4K1P1/PP5P/8 b - - 0 29
Fritz wants to play ...a4 but my basic instincts are saying that ...b6 may be better. I can't see the merit of ...a4 at all.

Desmond
22-06-2007, 10:58 PM
I think the benefit of ...a4 is that black will be left with reserve tempo to waste with the b-pawn if this is useful at some point in the future.

Kevin Bonham
22-06-2007, 11:03 PM
I think this was drawn and there may well have been many ways to do it.

I think your crucial error was ...b5. This move meant that it was no longer possible for you to turn your q-side 4-3 majority into a passed pawn situation without that pawn being an isolated d-pawn. And if that pawn is isolated then white's king covers it while white pushes his k-side majority (meaning that white has an outside majority) - at some point your king will have to divert to stop his majority and then his king cleans up your d-pawn and is closest to your a and b pawns.

Capablanca-Fan
22-06-2007, 11:51 PM
I think your crucial error was ...b5. This move meant that it was no longer possible for you to turn your q-side 4-3 majority into a passed pawn situation without that pawn being an isolated d-pawn.

Here I was thinking that it was an Open Ruy López gone wrong for Black, when this weakness was an unforced error in the endgame. There is something to be said for revising the advice in Capablanca's Chess Fundamentals, such as a general rule in pawn endgames: "advance the pawn which is unopposed". He analyses mistakes very much like that one.

Phil Bourke
23-06-2007, 12:02 AM
Thanks, for some reason during the game I had the idea that I would be able to play ...Kd6 and ...c5 activating the Qside, but alas had miscalculated White's play on the Kside. Thanks for the reasoning behind ...a4, will hopefully retain that for future use :) Plus I like that simple advance the pawn that is unnopposed rule as a guide too, should help me find better lines in future games. Thanks again fellows, I appreciate it greatly.

Capablanca-Fan
23-06-2007, 02:23 AM
1...h5 2.gxh5 Ke6 3.Kf2!! and black is lost; Black can not prevent White's king reaching g4.

Good one! Just like Alekhine – Yates, Hamburg 1910 (http://www.chessgames.com/perl/chessgame?gid=1011773).

Kevin Bonham
23-06-2007, 02:34 AM
Good one! Just like Alekhine – Yates, Hamburg 1910 (http://www.chessgames.com/perl/chessgame?gid=1011773).

Similar winning pattern of not approaching the two pawns was missed in a game between two c.1000 juniors in the Tas Open a few weeks back.

Capablanca-Fan
23-06-2007, 02:45 AM
One of Capa's illustrative games was against Salwe and allies (http://www.chessgames.com/perl/chessgame?gid=1316947), where Capa was told that the position was balanced, but Capa pointed out Black's backward c-pawn, meaning that White's e-pawn would be decisive, as it happened.

Fischer once got Reshevsky (http://www.chessgames.com/perl/chessgame?gid=1044252)into this sort of mess, but won the c-pawn and then finished quickly. Black's problems are not over even if he eliminates the backward c-pawn, because White's outside passer would often win, which is what I thought might have happened in this game.

It did happen in Em. Lasker – Rubinstein, St Petersburg 1914 (http://www.chessgames.com/perl/chessgame?gid=1119764), but, to Euwe's chagrin, not in his crucial game with Botvinnik in Groningen 1946 (http://www.chessgames.com/perl/chessgame?gid=1032141). Although it looked like a mirror image of Lasker – Rubinstein, the extra rook's pawns mean that the K+P ending would be drawn—Black can't just get the knight's pawn for nothing; he must go for the rook's pawn, but White gets the knight's pawn and queens in time.

Capablanca-Fan
23-06-2007, 02:52 AM
Similar winning pattern of not approaching the two pawns was missed in a game between two c.1000 juniors in the Tas Open a few weeks back.

And between players in the Wellington Chess Club in about 1995:

8/8/8/5p1k/5P1P/8/5PK1/8 w - - 0 29

Play went 1. Kf3 (?) Kxh4 2. Ke3? Kg4 3. f3+ Kg3=. In an article in New Zealand Chess, I pointed out that 1. Kg3 would force the BK back and gain a tempo, but even after 1. Kf3 Kxh4 2. Ke2! would win. Black can't stop White's K reaching e5, via d3 or d4, avoiding e3 unless Black plays ... Kg4. Note that White will always win if the "trébuchet" mutual zugzwang (http://en.wikipedia.org/wiki/Zugzwang#Tr.C3.A9buchet)with WKe5 v BKg4 is reached, because the back doubled pawn provides the reserve tempo if needed. Grigoriev composed a study about a very similar position in 1935.

Capablanca-Fan
23-06-2007, 10:23 AM
There is a drawing line for Black after ...h5 2 gh5 Ke6 but in this line White can win by 2 Kf3!

Good point. This "cooks" Kevin's elegant triangulation, because 2... Kf5 can be answered by 3. h6! gxh6 4. h5, and White reaches g4.


8/1p4pp/2p1k3/p2p4/3P1P2/4K1P1/PP5P/8 b - - 0 29
Fritz wants to play ...a4 but my basic instincts are saying that ...b6 may be better. I can't see the merit of ...a4 at all.

It could fix a target that would result in this pawn very close to queening. Another hint with an outside pawn majority in pawn endgames: it often pays to exchange pawns there, so when the defending K is drawn towards it, there are no targets to counterattack while the attacker cleans up the other side.

Kevin Bonham
23-06-2007, 01:38 PM
There may be more than one solution but I'm not convinced that 2.Kf3 is it!

(For those not familiar with "cook" in this context it is a problem reference - a problem is "cooked" if it has multiple solutions - in formal problem composition this is a bad thing, in trying to prove a position is a win it isn't. Jono isn't saying there's anything wrong with my 1...h5 2.gxh5 Ke6 3.Kf2 winning line, just that it might not be the only one.)

As far as I can tell 1...h5 2.Kf3 c5!? is drawn - 3.dxc5 (forced) Kxc5 and now:

(a) 4.gxh5 Kd4
(b) 4.f5 Kd6

and

(c) 4.g5 d4! (I think ...Kd6 actually loses in a rather drawn-out fashion) 5.f5 Kd5

appear to all be draws.

Corrections welcome, I am not 100% about these lines.

Capablanca-Fan
23-06-2007, 02:09 PM
There may be more than one solution but I'm not convinced that 2.Kf3 is it!

Oh, sorry, I meant after 1... h5 2. gxh5 Ke6 -- then 3. Kf3 instead of the perfectly good Kf2 triangulation; .... Kf5 4. h6!.


(For those not familiar with "cook" in this context it is a problem reference - a problem is "cooked" if it has multiple solutions - in formal problem composition this is a bad thing, in trying to prove a position is a win it isn't. Jono isn't saying there's anything wrong with my 1...h5 2.gxh5 Ke6 3.Kf2 winning line, just that it might not be the only one.)

Exactly.

Trent Parker
23-06-2007, 02:13 PM
Anyone ready for another NSW OPEN U1600 Endgame Study?

8/1p2kp1p/1p1p2p1/3P4/P1P5/1P3K2/6PP/8 w - - 0 32

White to play

Kevin Bonham
23-06-2007, 03:17 PM
Oh, sorry, I meant after 1... h5 2. gxh5 Ke6 -- then 3. Kf3 instead of the perfectly good Kf2 triangulation; .... Kf5 4. h6!.

Yes I was looking at 2.Kf3 because that is what Phil's post said. But as far as I can tell 3.Kf3 is drawn too. After 3...Kf5 4.h6 gxh6 5.h5 (or black will take that square and stop white from making progress on the k-side) b4! 6.b3 (forced) Kf6 and white can again get to g4 but this time cannot do anything there as he has no way to waste a tempo and force black to go away. 7.Kg4 Ke6 8.f5+ Kf7 goes nowhere.

At the moment 2.gxh5 Ke6 3.Kf2 is still the only win I have found against 1...h5.

Kevin Bonham
23-06-2007, 03:52 PM
Anyone ready for another NSW OPEN U1600 Endgame Study?

Yikes, that one is difficult. Black has to play carefully to stop White from either cleaning up the kingside pawns or breaking through on the queenside but I think that black can do this. My initial bid for this one will be "draw".

MichaelBaron
23-06-2007, 04:51 PM
I agree with Kevin, it looks like a draw..

If white could get his pawn on g5 and get say...g5 pawn vs black's g and h pawns on kinside he could prob explore his queenside advantage. But it appears to be highly impossible.

Rincewind
23-06-2007, 05:09 PM
Yikes, that one is difficult. Black has to play carefully to stop White from either cleaning up the kingside pawns or breaking through on the queenside but I think that black can do this. My initial bid for this one will be "draw".

I think it can be drawn too. In some of the lines I considered it is white who has to play carefully. This is mainly because black's chances of making a passed pawn on the kingside look better than white's chances on the queenside. Mind you none of these lines were explored on a board so perhaps there is a way for white to press the advantage.

eclectic
23-06-2007, 05:35 PM
can such positions be subthreaded here to facilitate continuity of analysis exchange?

Capablanca-Fan
23-06-2007, 06:05 PM
In some of the lines I considered it is white who has to play carefully. This is mainly because black's chances of making a passed pawn on the kingside look better than white's chances on the queenside.

That's important to note. Doubled pawns can be more effective than straight ones when facing a pawn majority. Sure, these ones would be vulnerable to Kb5, but White will be too busy keeping an eye on the K-side majority. White definitely needs to avoid swapping that down from 3v2 to 1v0.

Capablanca-Fan
23-06-2007, 06:06 PM
Yes I was looking at 2.Kf3 because that is what Phil's post said. But as far as I can tell 3.Kf3 is drawn too. After 3...Kf5 4.h6 gxh6 5.h5 (or black will take that square and stop white from making progress on the k-side) b4! 6.b3 (forced) Kf6 and white can again get to g4 but this time cannot do anything there as he has no way to waste a tempo and force black to go away. 7.Kg4 Ke6 8.f5+ Kf7 goes nowhere.

At the moment 2.gxh5 Ke6 3.Kf2 is still the only win I have found against 1...h5.

Right you are. So the triangulation is not just the most principled win, but the only win.

Phil Bourke
23-06-2007, 07:47 PM
Right you are. So the triangulation is not just the most principled win, but the only win.
As anyone cracked the example given by TCN for more than a draw, I can't see how Black makes any headway at all in this position.

Kevin Bonham
23-06-2007, 08:51 PM
can such positions be subthreaded here to facilitate continuity of analysis exchange?

You can do this yourself using the Display Modes function (either Hybrid or Threaded mode will do it) although it takes a bit of getting used to (and some posts end up on the wrong side if people click reply to one and then comment on the other).

eclectic
23-06-2007, 08:57 PM
You can do this yourself using the Display Modes function (either Hybrid or Threaded mode will do it) although it takes a bit of getting used to (and some posts end up on the wrong side if people click reply to one and then comment on the other).

i forgot about that

thx

Rincewind
23-06-2007, 09:11 PM
As anyone cracked the example given by TCN for more than a draw, I can't see how Black makes any headway at all in this position.

I don't think either can win but either can lose if they try too hard.

The thing is White cannot make a passed pawn on the queen-side without his king and if his king abandons the king-side then black makes a passer which probably wins for black.

White wants to lock up the king-side so as to be able to go over to the queen-side but I don't think that is possible by force. So therefore draw.

Does anyone have a plan that might work?

[Edit: I've not looked at it closely but have the impression that if the doubled pawns were on the kingside instead of the queen-side then it probably would be winning for white. e.g. Move b7 pawn to c7 or a7 and f7 pawn to g7. Now the plan to lock up the king-side and create a passed pawn on the queen-side might be realisable.]