< Earlier Kibitzing · PAGE 2 OF 2 ·
|Jun-18-05|| ||patzer2: Today's puzzle solution is the subtle 19. g5!! which is a deflection that sets up a winning pinning combination, eleven moves deep. The combination is so difficult that, even after 25 minutes of calculation at 2 GHZ and 16 depth, my strong chess program hasn't yet computed a winning advantage (overlooking the winning followup 20. Bh3! in favor of 20. fxg5 ).|
|Jun-18-05|| ||EinZweiDrei: God, seeing this sort of genius come from Tal never fails to destroy the very fiber of my life.|
|Jun-18-05|| ||melianis: Thank you <patzer2> <al wazir> for the analyses. In the game, 28...Rf8?! intending f3 if the h3-rook leaves the 3rd file, but after 28...Rf8 29.Bxf7 Rxf7 30.Rd3 Bc5 31.Kf3 ... black is lost also here after few more moves.|
Very hard to see the bishop manouvre giving time for the rook.
|Jun-18-05|| ||vonKrolock: <patzer2> For me the sequence as played actually <18.de> etc would be prefered by an attacking master even without calculation of long variations, just because it combinate the idea of ♗e6 followed by ♖f3 with an imediate threat over 'd7' - thus winning time for the assault: in comparison, 20.♗h3 allowing the key defensive move 20...♖ae8! is much slower: as You pointed out, white would have to play for the endgame with his ♗♗s and better ♙♙'s shape|
<20...ef> Another defensive idea: 20...g4 (better than 20...e4 21.♗e6 , and if ♔h7 then 22.♕h3 followed by the exchange of ♕♕s and f4xg5) 21.♗g4 e4!? (zk) and if 22.♗e5, then ♔h7 or ♔h8 etc – now at least black avoids further opening of lines – the vertical ‘f’- and a fast ♖f3-h3 ... Of course that white would then have still the better prospects in the final with his ♙-plus (‘e4’ will fall in almost all reasonable lines) – but with still a fair fight and some chances for black in ♖♖s and oposite coloured ♗♗’s <♘ote: considering quite probable the exchange ♗b2 vs ♘d7> possible endgames – in some situations, the SPLAWM ( Splitted Pawn Majority - f4-h2 vs g6 - can becomes a factor favouring equilibrium
|Jun-18-05|| ||vonKrolock: <in comparison, 20.Bh3 allowing the key defensive move 20...Rae8! is much slower:> better interpolating : <18.g5 hg 19.fe de> between <in comparison,> and <20.Bh3 aollowing >|
|Jun-18-05|| ||patzer2: After 2 hours and 30 minutes of calculating White's 20th move, to 17/49 depth and 1272kN/s my Chess program gives:|
M Tal - R Meulders
r4rk1/pp1n2q1/2p2bp1/4p1p1/5P2/1P1Q4/PBP3BP/4RRK1 w - - 0 1
Analysis by Fritz 8:
1. ± (1.09): 20.fxg5 Bxg5 21.Bh3 Rad8 22.Be6+ Kh7 23.Qh3+ Bh6 24.Kg2 Qe7 25.Bc1 g5 26.Rxf8 Nxf8 27.Rxe5 Qf6
2. (0.53): 20.fxe5 Bxe5 21.Bd5+ cxd5 22.Qxd5+ Kh7 23.Bxe5 Nxe5 24.Qxe5 Rxf1+ 25.Kxf1 Qf7+ 26.Kg2 Rf8 27.Qe7 Kh6 28.Qxf7
(0.44): 20.Bh3 exf4 21.Be6+ Kh8 22.Rf3 g4 23.Bxg4 Bxb2 24.Rh3+ Kg8 25.Be6+ Rf7 26.Qxd7 Qd4+
4. (0.28): 20.Qc4+ Kh8 21.fxg5 Bxg5
5. = (0.00): 20.f5 e4 21.Qc4+ Kh8 22.Bxf6 Nxf6 23.Bxe4 Nxe4 24.Qxe4 gxf5 25.Rxf5 Rxf5 26.Qxf5 Rf8 27.Qg4 Qf7
6. = (0.00): 20.Ba3 Rf7 21.fxe5 Bxe5 22.Bh3 Re8 23.Bd6 Rxf1+ 24.Rxf1 Bxd6 25.Bxd7 Qxd7 26.Qxg6+ Kh8 27.Qh6+
7. (-0.34): 20.Bxe5 Bxe5 21.fxe5 Rxf1+ 22.Bxf1 Nxe5 23.Qe3 Re8
Either Tal or the computer is missing something here, and I'm guessing it's probably not Tal.
|Jun-18-05|| ||Hemmeireoid5: I had a look at g5 and didnt get much further. Then I saw that it Was a certain M Tal that was playing and then I knew that I didnt have to look any further. I dont think Tal calculated to the end of the game either, I think he simply looked at the position and saw that he had an initiative for his pawn and that at some point, both his rooks would come into the game and that Black would be an exchange down. Unfortunately I didnt see it at all. Typical Tal, for a player like me g4 and f4 seems too early before bringing in the rook on a1, but Tals middle games are impecable and I would say that if he calculated anything at all it was in the middle game. After g5 I think the furthest he calculated was 25 Be6+ which to be honest is enough.|
|Jun-18-05|| ||kevin86: Is it my imagination that Tal seems to be favored at this site-specifically in sharp combinations.|
My favorite in this aspect of chess has always been Morphy. Am I just so cold-war minded or is it just the fact that I haven't seen many of Tal's games except here?
|Jun-18-05|| ||aw1988: Wow, Fritz doesn't find it? I wonder what Sunday will be like...|
Granted, computers are horrible in strategy and long-term planning, but this is a tactical position!
|Jun-18-05|| ||Eric Xanthus: (If anyone still cares) Was there any urgency to black's 28..Be3? At several points in the subsequent defense, it seems that 28..c5 would have been very helpful. Can someone enlighten me?|
|Jun-18-05|| ||aginis: <patzer2><
1. ± (1.09): 20.fxg5 Bxg5 21.Bh3 Rad8 22.Be6+ Kh7 23.Qh3+ Bh6> How did the bishop get from f6 to h6?? the only legal move from here is 23...Qh6 and then 24...Qe7 makes no sense please clarify!! < 24.Kg2 Qe7 25.Bc1 g5 26.Rxf8 Nxf8 27.Rxe5 Qf6>|
< (0.44): 20.Bh3 exf4 21.Be6+ Kh8 22.Rf3 g4 23.Bxg4 Bxb2 24.Rh3+ Kg8 25.Be6+ Rf7 26.Qxd7 Qd4+>
This is in fact what was played. How does fritz analyse the position after
27.Qxd4 Bxd4 28.Kg2? IMHO black is up at least a full pawn, and positional considerations do not seem to counter the fact that black will lose another pawn on the kingside in a few moves. For example,
28...Rf8 29.Rd1 c5 30.c3 Be3 31.Rd7 b5 32.BxR+ RxB 33.RxR KxR 34.Rh7+ Kf6 Rxa7
or 28...Rf8 29.Rd1 Be3 31.Rd7 b5 32.BxR+ RxB 33.RxR KxR 34.Rh7+ Kf6 35.Rc7
i think fritz just ran out of steam in giving this variation a mere 0.44
|Jun-18-05|| ||aginis: <EricXanthus> black plays 28...Be3 in order to be able to play 29...Re8; If 28...Re8?? 29.Bxf7+ Kxf7 30.Rh7+ Kf8 31.RxR (the exchange which Be3 is meant to prevent.) 31...KxR 32.Rxb7 |
|Jun-18-05|| ||Eric Xanthus: <aginis> I understand that you play 28..Be3 to play 29..Re8, but that just delays my question a bit. Why not interpose 28..c5? It seems that this is the last point in the continuation where black will have time for this move, and there doesn't seem to be an immediate need for 29..Re8.|
|Jun-18-05|| ||korger: On my 2.8 GHz Pentium 4 under Linux, GNU Chess finds the winning solution 19. g5! in 433 seconds.|
I think GNU Chess deserves more credit than it usually gets, but there's a problem with it that is particularly vexing for analysis: its Principal-Variation lines are useless beyond depth 6 or so, because for some weird reason it reports only the first 1 or 2 plies. Usually the longest lines are given for depth 4 (which are useless due to lack of depth), and then they gradually diminish, to finish the analysis like this:
Ply Time Eval Nodes Principal-Variation
8& 6.50 143 2990725 Ba3 Rf7
8. 18.04 143 8764118 Ba3
9& 35.06 165 15450660 Ba3 Rf7
9. 65.19 165 29702765 Ba3
10& 139.09 136 61809545 Ba3 Rf7
10& 432.92 160 203714523 g5 hxg5
10. 469.77 160 220955629 g5
Not very informative. Does anyone know why it doesn't provide more moves, when it can obviously see them?!
(Ehm, the output looks quite bad, since subsequent spaces are contracted in the message. What a pity.)
|Jun-18-05|| ||ThomYorke: g5! is such a good move. It allowed the light squared bishop to go to a marvelous position.|
|Jun-18-05|| ||aginis: <black knight c6>$<jahhaj>
20. ... Rad8 21.Be6+ Kh8 22.Rf3 etc.
|Jun-18-05|| ||aginis: <EricXanthus> sorry i forgot to point out that c5 only works after the rook no longer protects c3 otherwise
28...c5 29.c3 Be3 etc|
|Jun-18-05|| ||Eric Xanthus: Thanks <aginis>. I lost track of that rook on h3 and planned a simple capture if 29.c3. Too bad; I thought I might be on to a defensive resource. :(|
|Jun-18-05|| ||Boomie: 20...exf4 seems to lose in all variations. 20...e4 survives to the endgame. White still has a positional advantage but no greater than before 19. g5.|
20... e4 21. ♗e6+ ♔h8 22. ♕h3+ ♕h7 23. ♕xh7+
1. (23. ♗a3 ♕xh3 24. ♗xh3 g4
1a. (24...♗d4+ 25. ♔h1 ♖fd8 26. fxg5 (1.82) D=15)
25. ♗xg4 ♗d4+ 26. ♔g2 ♗c5 27. ♗b2+ ♘f6 28. ♖xe4 ♔h7 29. ♗xf6 ♖xf6 (0.90) D=14)
23... ♔xh7 24. ♗xf6 ♘xf6 25. fxg5 ♖ae8 (1.11) D=16)
|Jun-19-05|| ||Richard Taylor: This isn't very brilliant for a Tal game - g5 an h3 is/are pretty obvious (mind you I didn't try to work this one out! - just played the game over -lol) -of course it's handling the complexities after that that count: but as an idea it's fairly straightforward- that's not to say Tal didn't play excellently - I mean it isn't like some of his very complex and imaginative games where the position looks like a mad explosion...|
|Jun-20-05|| ||vonKrolock: in Fritz' line quoted above 20.fxg5 ♗xg5 21.♗h3 ♖ad8 22.♗e6+ ♔h7 23.♕h3+ ♗h6 24.♔g2 this last move seems somewhat obscure (maybe misterious) - but is seems that 'The ♔ing' agree with his coleague: it's a move that arithmetically will lead to the winning for white of a ♙♙s plus in manobrating play - but i sincerelly doubt that a human player would chose 24.♔g2 in OTB or Solving Chess conditions...
Well, if Tal played HIS own way and won, and the artificial masters are discording - it's a good sign |
|Jun-23-05|| ||patzer2: <aginis: <patzer2>< 1. ± (1.09): 20.fxg5 Bxg5 21.Bh3 Rad8 22.Be6+ Kh7 23.Qh3+ Bh6> How did the bishop get from f6 to h6?? the only legal move from here is 23...Qh6 and then 24...Qe7 makes no sense please clarify!!> After 20...Bxg5 in this Fritz 8 line, the Bishop is in position to move from g5 to h6 with the ensuing 23...Bh6.|
Note that this line was clipped (not read and typed) and pasted from the Fritz 8 analysis bar.
|Jun-23-05|| ||patzer2: < (0.44): 20.Bh3 exf4 21.Be6+ Kh8 22.Rf3 g4 23.Bxg4 Bxb2 24.Rh3+ Kg8 25.Be6+ Rf7 26.Qxd7 Qd4+ This is in fact what was played...
I think fritz just ran out of steam in giving this variation a mere +0.44>.
|Jun-23-05|| ||patzer2: Apparently Fritz 8 initially assessed that Black could improve over 28...Be3 in this game with 28...Raf8!?|
However, after 28...Raf8!? 29. Rd1! Be3 30. Rd7 b5 30. c4! a6 31. Rd6 bxc4 32. Bxc4 Kg7 33. Bxf7 Rxf7 34. Rxc6+ White retains a winning advantage.
|Jul-12-05|| ||korger: After giving a 1 meganode hash table to GNUChess 5.05 (yep, I found this was the solution to my problem above), the program finds 19. g5 in 140 seconds, and gives the variation|
19. g5 hxg5
20. fxg5 ♗xg5
21. ♗h3 ♖ad8
22. ♗e6+ ♔h7
23. ♕h3+ ♗h6
24. ♖xf8 ♘xf8
25. ♗xe5 ♕e7
with a score 1.14 pawns in favor of White. However, a quarter an hour later it opts for
19. f5 gxf5
20. ♖xf5 e4
21. ♕c4+ ♔h8
22. ♗xf6 ♘xf6
23. ♗xe4 ♘xg4
and the score is 1.22. I agree that the difference is not significant, but this would say that 19. f5 is a viable alternative to 19. g5. The fact that the computer's g5 line is not the same as Tal's is undeniable (already differs at the 20th move), but in this case I tend to think the computer is right. This is a complicated tactical position, and nobody beats a comp there! But I wonder why the other chess engines don't come up with 19. f5. The variations above make me think it's even better, after all.
< Earlier Kibitzing · PAGE 2 OF 2 ·