< Earlier Kibitzing · PAGE 2 OF 2 ·
|Jan-08-05|| ||tamar: <gypsy> What if Loyd had played 44 Kg5 with the idea of protecting the g pawn? White can get his King to f4 by force with: |
44 Kg5 Ke5 45 Kh6 Kf4 47 Kh5
but then advancing the f pawn with 47... f5 is bad because on
48 g5 White is quicker.
|Jan-09-05|| ||Gypsy: Congratulations <tamar>!! Your 44.Kg5!! -- I am convinced of it now -- is a refutation of the Black's breakthough.|
The most slipery variation I see as still to be played is 44.Kg5!! Ke5 45.Kh6 Kf4 46.Kh5 f6!? 47.Kh4 Kf3!? 48.Kh5! (48.Kh3? Ke4 ) Kf4 (48...Kg3? 49.g5! ) 49.Kh4 Kf3 =
I am in awe that you found the 44.Kg5, it did not even cross my mind. Perhaps I could have found it if somebody gave me the position after the 43...Kxe4 as a puzzle. But in view of the flow of the last few moves, the reversal Kf4xg5-h6-g4-h6 was next to impossible for me to see.
|Jan-09-05|| ||tamar: <gypsy> I was intrigued Sam Loyd played at Paris 1867, and puzzled that he would be outplayed in a simple King and Pawn endgame, so I just examined all the variations you gave and then the alternative legal moves. 44 Kg5 is much easier to consider when you know for sure 44 Kg7 loses to 44...f5!, and 44 g5 Kf4 45 Kh5 Kg3! |
Moving in front of the pawn is so anti-intuitive that psychologically, it gets discarded unless there is no other way. Just out of curiousity, I also looked at 44 Kh7, which of course loses, but I never got around to looking at the only other legal move 44 Kh5, which draws also, I think.
Now the question is: where did Winawer slip up?
|Jan-09-05|| ||beatgiant: <tamar: Now the question is: where did Winawer slip up?>|
Given what we now know about the pawn ending, 36...b5 allows a draw. I suggest 36...g5+ 37. hxg5 hxg5+ 38. Ke3 Ke5 39. Ra1 first, and only now 39...b5 40. axb5 cxb5 41. Rxa5? bxc4 42. Rxc5 dxc5 43. bxc4 f6, with zugzwang.
|Jan-09-05|| ||fred lennox: For me, 36.Ra2 makes the rook inexcusably passive. So, offhand, I'd say attack where the rook isn't. Instead of 36....b5 maybe...f5. 37.gxe5...gxe5 38.dxe5...Rxe5+, king moves then...d5 gives black a passed pawn. |
|Jan-09-05|| ||tamar: <beatgiant> <I suggest 36...g5+ 37. hxg5 hxg5+ 38. Ke3 Ke5 39. Ra1 first, and only now 39...b5 40. axb5 cxb5 41. Rxa5? bxc4 42. Rxc5 dxc5 43. bxc4 f6, with zugzwang.> Yes that would be a clear win, but earlier what if White did not agree to exchange rooks in your line
(39 Ra1 b5) and instead played 39 Rf2, isn't the win in doubt as the rook becomes active on the f and h files? |
<fred lennox> I would give 36 Ra2 an exclam, even though it looks inexcusably passive, as you say. It does the dirty work of preventing the break with 36...b5 which was obviously Winawer's intention. The immediate breaks other than ...b5 don't look good enough yet, primarily because White can exploit the clumsy position of the rook on c5. I haven't looked at it closely, but running the g pawn looks very appealing against the f5 break.
36...f5 37 exf5 gxf5 38 g5! hxg5 39 hxg5
and if 39...Kf7 Rd2, and if 39...Re5
|Jan-09-05|| ||Gypsy: <Now the question is: where did Winawer slip up?> |
To me, Winawer's whole play between <29...Kg7> and <36...b5> is less than convincing. Given the x-ray on the e-file, <f5> by Black somewhere before
<31...Kf6(?)> would cut through the chase and either captured a second pawn or liquidated into an a won pawn-endgame. Prototypical variations go as follows:
29...f5(!) 30.gxf5 gxf5 31.exf5 Rxe1 32.Kxe1 Kf7 ...
29...f5(!) 30.gxf5 gxf5 31.Kf3 Kf7 32.Re3 Kf6 33.exf5 Rxe3+ ...
|Jan-09-05|| ||Gypsy: In the earlier part of the game, I realy like Winawer's maneuver <13...Bxe4 14.dxe4 Bf6 15.Qd2 Bxb2 16.Qxb2 Qf6! 17.Qd2 Ne5> which gives Black the upper hand in the game. The maneuver is of course based on a small tactical trick, namely that 17.Qxf6?? looses a piece to 17...Nxe2+.|
Of course, <18.Rad1?> is a tactical error that looses a pawn.
|Jan-09-05|| ||tamar: This is the first game of Winawer I have really studied. Not a bad player. I just noticed that one of the points of 17...Ne5! is that 18 Nxe5 is met not with 18...dxe5 or even with the better 18...Qxe5, but with 18...Qg5!! mating or winning the queen. |
|Jan-09-05|| ||Gypsy: <...18 Nxe5 is met not with 18...dxe5 or even with the better 18...Qxe5, but with 18...Qg5!! mating or winning the queen. > Thanks for pointing that out! The two black knights are a terrible menace. It seems that the only way to hold material was to play 18.Qd1. But what a humiliating tally of the last five or so moves. Hindsight 20/20; in view of the Qg5, 17.Qc2 was probably preferable to <17.Qd2>. Better yet, 15.Bxf6 was preferable to <15.Qd2>; but I suspect Loyd did not think <16...Qf6!> playable, untill too late. |
|Jan-09-05|| ||beatgiant: <tamar>
<what if White did not agree to exchange rooks in your line (39 Ra1 b5) and instead played 39 Rf2, isn't the win in doubt as the rook becomes active on the f and h files?>
I still suspect Black should be winning as of move 36. Perhaps I gave too many preparations and should try 36...g5+ 37. hxg5 hxg5+ 38. Ke3 b5 immediately?
|Jan-09-05|| ||tamar: <beatgiant> < 36...g5+ 37. hxg5 hxg5+ 38. Ke3 b5> 39 Kd4! might hold by just waiting. I am sure a win is there around 36th move too, but it may be a less forcing line than I am looking for. |
|Jan-09-05|| ||Gypsy: Before leting go of the <tamar's> defensive discovery -- 44.Kg5(Kh5)! -- I want to see if we can extract a lesson from it on a strategic or meta-strategic level. (We know it works on the tactical level.)|
How can such <running in place> save White in the end-game when his thematic continuations (Kg7 and/or g5) both loose? My suspicion is that here we have a strong case of <Tartakower's maxim> about a threat being frequently more powerful than its execution.
Breaking the position after <43...Kxe5> into its elements, we can see (i) the pair of Q-side blocked pawns c4/c5, which are dead even; (ii) the pair K-side neighboring pawns g4/f7, where White holds a slight but significant tempo-advantage; and (iii) the position of the kings, where Black holds a huge advantage. Black king is well positioned to attack the Q-side as well as to defend the K-side. In contrast, for all practical purposes, White king's play is now restricted to the K-side.
There is one more key feature of the position: (iv) if Black decides to immeditely recklesly attack on the Q-side, then White gets through on the K-side faster! And this is the key: Black king can not leave the K-side untill the tension there is released. So, while White attacking strategies (44.Kg7 and 44.g5) both release the tension but fail to win/draw for tactical reasons, tamar's attack-defering strategy keeps the tension up and never lets Black king part to execute his evil dead against the c4-pawn.
Chess is a logical game, after all! :-)
<gentlemen> It has been a pleasure and priviledge to study this game with you.
|Jan-09-05|| ||fred lennox: <tamar> your right. the Ra2 move was a good move though as it turns out, he is dancing to blacks tune. Maybe the exchange of queens would of been in whites favor 17.Qxf6. The queen seems rather passive after 17.Qd2 |
|Jan-09-05|| ||Gypsy: <Maybe the exchange of queens would of been in whites favor 17.Qxf6> Alas, 17.Qxf6? Nxe2+. |
|Jan-10-05|| ||beatgiant: <tamar>:
<I am sure a win is there around 36th move too, but it may be a less forcing line than I am looking for.>
The simple 36...d5 may be good enough. For example, 36...d5 37. exd5+ cxd5 38. Re2+ Kd6 39. Rd2 Kc6 40. cxd5+ Rxd5 41. Rc2+ Kd6 42. Ke3 Rd1 43. Rf2 Ke6, and I don't see how White can defend against the double threat of ...Rb1 and ...Rh1.
|Jan-10-05|| ||tamar: <beatgiant> Nice. 36...d5 is the move Winawer should have played. I hadn't considered playing it immediately, only after 36...g5+ 37 hxg5 hxg5+ 38 Ke3 d5 39 Kd4! b6 40 Re2+ Kd6 41 Rf2 with White getting active again. But 36...d5 leaves the King on f4, which blocks the f file, and also eliminates the saving Kd4 move. |
|Jan-10-05|| ||Gypsy: <bg, tamar, fl> After 36...d5, I would try to set up White defenses on the e- and mainly f-file. I do not know if it ultimately succeeds to hold the position, but here is a line as 'food for thought'.|
36...d5 37.exd5 cxd5 38.Re2+ Kd6 39.cxd5 Rxd5 40.Re3 and, to fix ideas, say, 40...Rd4+ 41.Kg3 Rb4 42.Rf3 ...
Such a defensive setup could be hard to crack.
|Jan-10-05|| ||Gypsy: <beatgiant> You seem to be fond of rook endgames. Here is good one. Rubinstein vs Duras, 1912 |
|Jan-10-05|| ||Gypsy: 29...f5(!) 30.gxf5 gxf5 31.Rg1+(!) keeping the rooks on must be the toughest defense here 31...Kf7 32.exf5 Rxf5+ 33.Ke3 (g-file is out of limits). |
This brings up a challenging but it would seem winnable rook-endgame. Pawns are still on both wings and, on the Q-side, White can no longer set up his optimal defensive structure.
|Jan-10-05|| ||tamar: <36...d5 37.exd5 cxd5 38.Re2+ Kd6 39.cxd5 Rxd5 40.Re3 and, to fix ideas, say, 40...Rd4+ 41.Kg3 Rb4 42.Rf3 ...> |
<gypsy> Since the f pawn comes under attack at the end of this line, I offer as an improvement 41...f5 42 gxf5 gxf5. At this juncture, White has to make a decision, to stand pat with 43 Rf3 which allows Black to cross over the e file with his King, or to play 43 Re8 and pick a pawn target and/or harass the Black King. I don't see a clear-cut win against the active line, but overall I think 41...f5 would avoid the inertia Black faces in the 41...Rb4 42 Rf3 line.
|Jan-10-05|| ||beatgiant: <tamar>
<I don't see a clear-cut win against the active line>
Your active line does look like a clear-cut win to me.
After 36...d5 37.exd5 cxd5 38.Re2+ Kd6 39.cxd5 Rxd5 40.Re3 Rd4+ 41.Kg3 f5 42.gxf5 gxf5 43.Re8, Black plays <43...Rg4+>, and now either 44.Kh3 Rb4 45.Ra8 Rxb3+ 46.Kg2 Ke5 47.Rxa5+ Kf4 wins on the kingside, or 44.Kf3 Rxh4 45.Rd8+ Kc5 46.Rf8 Rh3+, followed by 47...Rxb3 wins on the queenside.
|Jan-10-05|| ||Gypsy: <tamar/beatgiant> Indeed the 41...f5 looks like the right plan for the situation; it is a strong move and leaves White with some desperate choices.|
(This refers to the variation 36...d5 37.exd5 cxd5 38.Re2+ Kd6 39.cxd5 Rxd5 40.Re3 Rd4+ 41.Kg3 f5 42.gxf5 gxf5 ... from the game.)
|Jan-10-05|| ||tamar: <beatgiant> I based my hedging on the saying, 'Rook endings are never won' but Black is getting close. I find rook endings incredibly difficult to assess, perhaps because I don't have the certainty of which positions are theoretical wins. |
|Jan-11-05|| ||Gypsy: <tamar/beatgiant> <...I find rook endings incredibly difficult to assess...> That Rubinstein-Duras game listed above is a perfect example of how slippery rook endgames can be; and how beatiful. Comes to think of it, the slippery quality of rook endgames makes them quite fun to play -- they are a terror to anotate though. |
With this, I feel I still owe Mr. Winawer a few moves to realy set him off on a winning path in the line I have been reccomending:
<29...f5 30.gxf5 gxf5 31.Rg1+> (this is a good tactical excuse from the otherwise hopeless pawn endgame after exchanges on e4) <Kf7 32.exf5 Rxf5+ 33.Ke3> and now I found that the transparent <33...Rh5 34.Rg3> (34.Rf1+ Ke7(Ke6) only helps Black) <34...Rh4> is preferable to all the fancy convoluted maneuvers I tried. This way the Black rook asserts good activity and also grabs space by constricting most of the mobility of White pieces. They now trip over each other and, with a little caution, there is gererally no good active play for them. Black now shall mobilize his Q-side mayority and press on towards the goal line. Black king should stay put while the 7-th rank needs protection, but start pushing the pawns from behind once they clear out.
< Earlier Kibitzing · PAGE 2 OF 2 ·