< Earlier Kibitzing · PAGE 2 OF 3 ·
|Jan-31-08|| ||goldfarbdj: I looked at Qxg2+ first, but saw that it didn't go anywhere. I blush to admit it, but I didn't see that Nxf3+ was met by Qxf3; but even without noticing that, I still found that e2 allowed all kinds of bad nastiness.|
|Jan-31-08|| ||Kwesi: Hmmm, got it surprisingly quickly - tried Nf3+ first but when i saw it failed to Qxf3, e2! suggested itself...|
|Jan-31-08|| ||ahmadov: <Kwesi: Hmmm, got it surprisingly quickly - tried Nf3+ first but when i saw it failed to Qxf3, e2! suggested itself...> The same here... Either we are having easy puzzles this week or this must be my week...|
|Jan-31-08|| ||zooter: <ahmadov: <Kwesi: Hmmm, got it surprisingly quickly - tried Nf3+ first but when i saw it failed to Qxf3, e2! suggested itself...> The same here... Either we are having easy puzzles this week or this must be my week...>|
Or rather i think this is the week of the advanced passed pawn....must watch out closely tomorrow for some similar "trick"
|Jan-31-08|| ||TheaN: 4/4
At least... if I can claim full points without seeing all the madness calculations by RV and MAJ. :).
28....e2 wraps White up. Threatening 29....exd1=Q and Nf3+, White should wake up at this point. Leaving e2 fails unforgivening:
29.Bxe5 exd1=Q 30.Rxg7+ Qxg7
29.Qd2 Nf3+ 30.Kf2 Nxd2 31.Rxg7+ Qxg7
29.Qa1 Nf3+ 30.Kf2 Nxd4 31.Qxd4 (Rxg7 Qxg7+ ) Qxg2+ 32.Kxg2 Bxd4
29.Qb3 (or anywhere else) Nf3+ 30.Kf2 Nxd4
Onto the captures now.
29.Rgxe2 Nf3+ 30.Kf1 Rxe2 31.Qxe2 (Bxg7 Rxe1+ ) Nxd4
29.Qxe2 Nf3+ 30.Qxf3 (Kf2 Rxe2+ 31.Rxe2 Nxd4 / Kh1 Rxe2 31.Rxg7 Qxg7 32.Bxg7 Rxe1+ ) Rxe1+ 31.Kf2 Rf1+ 32.Kxf1 Qxf3+ 33.Nf2 Rc7 with 34....Kh8 )
29.Rexe2 Nf3+ 30.Kf1 Nxd4 and here I thought not noticing Rxg7+ being so strong for White: 31.Rxg7+ Kxg7 (Qxg7 32.Rg1) 32.Rxe8 where suddenly Black seems to be stuck.
32....Rxe8 33.Qg4+ Kf8 34.Qxd4
32....Qh1+ 33.Ng1 (missed this one as well having looked at 31.Rxg7+) Rxe8 34.Qg4+ with Qxd4 and at most for Black. Larsen has thrown the towel in too fast.
|Jan-31-08|| ||Domdaniel: <MAJ> 8-core, eh? Just seeing the words "quick" and "20-ply" in the same sentence gives me a jolt. Silicon monster, indeed -- hope you have fun together.|
|Jan-31-08|| ||zb2cr: <MostlyAverageJoe>,|
8-core? Whooee...if that's a personal machine, buddy, you are rolling in the dough; if it's your boss's, I'll never tell!
|Jan-31-08|| ||DarthStapler: I don't get it|
|Jan-31-08|| ||Halldor: Got the main idea rather rather fast, 28... e2 (breaking the communications?), but then I spent a long time on all variations I could find, using board and pieces and writing down the moves. Think that's a good practice.|
|Jan-31-08|| ||Refused: Usually can't stand those "piece of cake" "less than 10 seconds" stuff, but here I got the solution for a Tal position rather quick (usually Tal is way above me).|
The obvious thing you want to play as black is Nf3+, but here it is simply not working right away. But the position after e2! is a different story, here I just followed Tal's advice ("Many sacrifices donít require concrete calculation at all. It is sufficient to only glance at the arising position to convince us that the sacrifice is correct.") and the arising position looked good enough.
The Queen can't capture on e2 because e1 is too weak. Any Rook capture finally allows Nf3+ and to leave the pawn on e2 is no option either.
Afterwards it was just working out some lines (just making sure it really works)
|Jan-31-08|| ||scholes: <"Many sacrifices donít require concrete calculation at all. It is sufficient to only glance at the arising position to convince us that the sacrifice is correct."> <refused> Tal was just joking otherwise he would have sacced a piece every game . And his glance is much different to glance of us mere mortals .|
|Jan-31-08|| ||ganstaman: <scholes: Tal won't had calculated this combination like ,okay there is problem Nxf3 line so try e2 blocking queen , rather he would have known the solution to this problem is giving a capture with check so e2 . These supergm's have to see these things 5-6 moves in advance , so it can't be done bruteforce>|
I honestly don't know if anything you said here is true. I'm sure Tal saw this all in advance, but I also think he likely tried Nf3+ in his mind first before realizing that he needed e2 before that.
|Jan-31-08|| ||ganstaman: <Halldor: Got the main idea rather rather fast, 28... e2 (breaking the communications?)>|
Another way to say "breaking the communications" that is often used in English when talking chess is "interference." But don't worry -- you'd be understood anyway.
|Jan-31-08|| ||kevin86: I was looking for a queen sac-like in the last two days,but I did find e2. If,instead,28...♘f3+ 29 ♕xf3 ♕xf3 30 ♖xg7+ ♔f8 31 ♖f1 ♕xf1+ 32 ♔xf1+ and white will win the e-pawn and will have a winning ending.|
|Jan-31-08|| ||muharrem: white should have played 27Rxg7+ and move on|
|Jan-31-08|| ||johnlspouge: Thusday (Medium): Black to play and win
Material: White is up a P. White has more space on the Q-side, but Black's Re8 supports the advanced passed Pe3 near the White Kg8, and lack's Ne5 and Qb7 are immediately available for K-side attack. White has the implicit beginnings of a K-side counterattack, with Bd4 potentially able to join the Rg2 in attacking the Bg7, possibly threatening after Ne5 moves 1.Rxg7 Kh8 [Kf8 2.Rf1#] 2.R moves +
Candidates (28...): Nf3+, e2, Qf3, Qxg2+
Let's sample 28...Nf3+.
28...Nf3+ 29.K moves Nxd4 30.Qg4 Kh8
Black is ahead, but Qd1 is out of the bottle and Ng5 creates some counterplay. Let's sample 28...e2 instead, to obstruct the Qd1 and to open the diagonal a7-g1.
28...e2 29.Rexe2 [Rgxe2 is inferior] Nf3+ 30.Kh1 [other K moves are no better] Nxd4 31.Rxe8+ [Rf2 Rf8, and the Rf2 is becoming a liability] Rxe8 and 32.Qg4 now looks positively weak because of 32...Re1+.
The disappearance of Rc8+Pe6 for Re1+Bd4 looks a lot better than leaving Rc8 and Re1 on the board. So, let's look at 28...e2 more closely. What are White's defenses?
29.Qxe2 Nf3+ 30.K moves [Qxf3 Qxf3 and White drops the Qe2 for a N and B] Rxe2 31.Rxg7+ Qxg7 and White's game is in ruins.
29.an innocuous Q move is not feasible, because it creates a monster passed Pe2 supported by Nd4 and surrenders f3 to the White Q. In essence, it is the same as the sample line for 28...e2 with Re8, Pe2, and Re1 still on the board.
Time to peek. I got the move 29...e2, but my logic felt shakier than usual. Time to check the kibitzing. I found the key move 28...e2 much less intuitive than many, apparently. <dzechiel>'s line
29.Qxe2 Nf3+ 30 Qxf3 Rxe1+ 31 Kf2 Rf1+ 32 Kxf1 Qxf3+ and 33...Qxh3
is a better defense than my sample from 28...Nf3+, but even that sample persuaded me to look for something better than 28...Nf3+.
|Jan-31-08|| ||zenpharaohs: Well I screwed that one up, my line was
28 ... Nf3+
29 Qxf3 Qxf3
which is actually fine except that white has the response
which I missed, but Rybka was able to show it to me immediately....
As it turns out black only has
28 ... e2
now it is safe to play
29 ... Nf3+
30 Kf1 Nxd4
white will play
(31 Rxe8+ is a loser)
and we have a choice, but Rybka advises
31 ... Kxg7
32 Rxe8 Rxe8
33 Qg4+ Kf7
34 Qxd4 Qh1+
35 Ng1 Qxh2
now white cranks up the windmill looking for perpetual check
36 Qd7+ Re7
37 Qf5+ Kg7
38 Qg4+ Kh6
but with care black can keep the game alive
39 d4 Rf7+
40 Nf3 Qh5
41 Qxh5 Kxh5
black has a small advantage (Rybka has it as -1.56) which might be enough to win but it's unclear. I will run this computation out quite a bit more before I know what to think.
But it looks like a bad resignation.
|Jan-31-08|| ||crwynn: <31. Rxg7+ Qxg7 32. Bxg7 Re1+ 33. Qxe1 Nxe1 34. Kxe1 Kxg7 35. Kd2 Kf6 36. Nf4 Ke7 37. Nd5+ Kd6 38. Nf6 Ke7 39. Nxh7 Rh8 40. Ng5 Rxh2+ 41. Kc3 Ra2 42. Kb3 Rd2 43. Kc3 Ra2 44. Kb3 Rd2 45. Kc3 Ra2 |
A quick 20-ply backslide shows no improvements for the black, hence 30... Rxe2 appears to be refuted.>
That is a sweeping conclusion to draw from basically useless computer analysis. Comp evaluations for pure rook vs. knight endings are often improperly skewed towards the player with the knight; it will give +0.50 or something for that side even though it is dead lost.
It also seems to have trouble playing such endings; obviously Black can easily avoid an exchange of h-pawns here and obviously he should do so, but computers want to allow it. I don't know if Black is winning with h-pawns on the board, but it is certainly difficult for White.
Black's plan is to play ...a5 at some point, which either opens the b-file or (after b5) allows the possibility Kc5-b4-a4. Keep h-pawns on the board on general principles - pawns on both sides of the board are desirable when you push for a win.
|Jan-31-08|| ||zenpharaohs: CRWynn: "Comp evaluations for pure rook vs. knight endings are often improperly skewed towards the player with the knight; it will give +0.50 or something for that side even though it is dead lost."|
I am not sure I agree with that. Can you give an example of a position where the player with the knight is dead lost? I can have a few engines look at it and see if they are mistaken.
|Jan-31-08|| ||crwynn: Spassky vs Korchnoi, 1977 if 48.h6 R1a6 49.Ra6 Ra6 had been played, I think Black is winning because of the d-pawn, but does your computer think so at first?|
|Jan-31-08|| ||zenpharaohs: CRWynn: "Spassky vs Korchnoi, 1977 if 48.h6 R1a6 49.Ra6 Ra6 had been played, I think Black is winning because of the d-pawn, but does your computer think so at first?"|
I don't get 48 ... R1a6 since there isn't a black rook at a1 in the position (they are at a2 and a7). But I have the game loaded up so let me know what the line is and I'll give you the numbers.
|Jan-31-08|| ||crwynn: right, I meant r2a6|
|Jan-31-08|| ||johnlspouge: <zenpharoahs>, Toga II 1.3.1 concurs that the resignation was bad.|
[plies 19/73 time 49:34 value -0.52]
29.Rexe2 Nf3+ 30.Kf1 Nxd4 31.Rxg7+ Kxg7 32.Rxe8 Qh1+ 33.Ng1 Rxe8 34.Qg4+ Kf7 35.Qf4+ Ke7 36.Qxd4 Rf8+ 37.Ke1 Qxh2 38.Ne2 Qd6 39.Qe3+ Kd7 40.c5 Qe6 41.Qd4+ Kc8 42.Qg7 Rf7 43.Qg2 Kb8 44.c6 Kc7 45.b5
|Jan-31-08|| ||zenpharaohs: OK in CRWynn: Spassky vs Korchnoi, 1977:
I think you must mean 48 ... R2a6 since 48 ... R7a6 loses.
After 48 h6 R2a6 49 Rxa6 Rxa6, the position is
click for larger view
Rybka has the score as -0.18 at depth of 28 plies. I've played it out a little. It looks like black will not promote the d pawn.
|Jan-31-08|| ||zenpharaohs: By the way if we are going to get deep into the endgame of the Spassky - Korchnoi we could consider putting the resulting kibbitzing under that game.|
< Earlier Kibitzing · PAGE 2 OF 3 ·