< Earlier Kibitzing · PAGE 3 OF 3 ·
|Jan-08-09|| ||zenpharaohs: Manic: <TheBish> "I think <zenpharaohs> line is Fritzed because an evaluation at the bottom of his post are given"|
I used Rybka 3 to get the values.
|Jan-08-09|| ||zenpharaohs: fosca: "Whew, now I don't feel like such a patzer for not being able to meet d6 satisfactorily and, therefore, rejecting d5. I settled on the very lame Rfe8 hoping Na5 would still be there....ha!"|
I ended up rejecting d5, and not finding anything to replace it with. I kept trying to figure out how to make Na5 work immediately. (Which it doesn't).
|Jan-08-09|| ||patzer2: While 25...Kf7 is decisive, even stronger might be 25...Bxb2 . If now 26. Rb1, then 26...Rc2 wins easily.|
|Jan-08-09|| ||LIFE Master AJ: I got this one right ... did it in around ten seconds too. |
Black could have resigned much ealier, perhaps he did not want to be the victim of a brilliant miniature.
|Jan-08-09|| ||LIFE Master AJ: Obviously no deep analysis is possible in such a short time, I was content that I saw 18...d5; with the idea of ...Na5, (winning material).|
|Jan-08-09|| ||Marmot PFL: With not much else happening in the position 18...d5! is a nice Sicilian shot, with the double threat dc4 and Na5.|
|Jan-08-09|| ||dukesterdog2: Took me a long time as I first was trying to find a sacrificial attack along the half open f file, but I eventually came up with 18...d5, and worked out the moves as played through 20...Na5 attacking white's queen and attacking the bishop twice.|
|Jan-08-09|| ||xrt999: <Breunor: Having seen the game, I think on 18 d5 19 exd5 Na5 20 d6, Black plays Qf7 similar to the actual game. Then if 21 Bxe6, for instance, Nxb3 22 Bxf7+ Rxf7 and black is up a piece.>|
This is the same exact line as the text, except in your line, black avoids the knight trade. In the text, after 24.Rad1, black has a bishop and knight to white's bishop, a ratio of 2 to 1.
In your line, black has 2 knights and a bishop to white's knight and bishop. You might think that this line is inferior because the ratio is 3 to 2, a smaller advantage than 2 to 1. Mathematically, the text would be theoretically correct, but after playing it through on CM, it is saying that they are virtually indistinguishable from each other.
[Just go to move 24 in the text and place a knight on f6 and c3]
|Jan-08-09|| ||Patriot: I didn't have time to analyze this enough but wanted to play 18...Na5 to win a piece but 19.Bxe6+ refutes it. So I looked at 18...d5 and if the bishop retreats then 19...d4 forks bishop and knight. I didn't get as far as looking at 19.Nxd5 which is a very nice attempt. For if 19...exd5 20.exd5 threatening both 21.dxc6 and 21.d6+. There are some really nice tactics here and with the main line it just continues to be sharp.|
|Jan-08-09|| ||xrt999: I dont know, after 24.Rad1 my patzer CM is only giving black -1.43 at a mere 10 plies. Black is up a knight but white is up 2 pawns, one being the isolated pawn on the 6th rank. This is totally playable as white. In other words, white has sacrificed a piece for a passed pawn. |
CM is telling me white starts to lose with 25.Bb6 (-2.5) whereas 25.Rd3 is maintaing the position at -1.4, but I dont really see this with my paltry 5 ply brain.
Can anyone find a win for white?
|Jan-08-09|| ||MostlyAverageJoe: <xrt999: I dont know, after 24.Rad1 my patzer CM is only giving black -1.43 at a mere 10 plies.>|
10 plies is completely useless on this puzzle. It would suffice on Tuesday. Let it run a bit more.
|Jan-08-09|| ||Patriot: <xrt999: I dont know, after 24.Rad1 my patzer CM is only giving black -1.43 at a mere 10 plies.>|
-1.43 may be close. I would calculate 3.25 (knight) - 2 pawns = 1.25 in black's favor. CM is probably adding a little more in black's favor because of white's isolated and weak pawn. Although 1 pawn is technically winning, it is still playable probably at the GM level and even more so below expert.
|Jan-08-09|| ||xrt999: I only run to 10 because I really dont grasp anything beyond that, and many times dont grasp 10 ply lines either, since I can only calculate maybe 5 or 6 moves deep. If I can grasp why a 10 ply line is winning, I consider it a total success. On the other hand, we are even, since I also consider <your> 20 ply lines completely useless, since they have absolutely no basis in real chess. |
Oops, let me rephrase that, what <I> consider real chess, OTB tourney play. I think what you and I consider "real" chess differs greatly. And thats completely fine, Im not saying that letting Hiarcs run overnight on a dual core P4 with infinity backslide analysis is not f'ing cool, because it totally is! Many times this analysis will show the most bizzare move, which seems totally wrong, and Hiarcs shows it winning. All you have to do is respond correctly and then make the next 19 moves correctly as well. Assuming we are in a complex middlegame, a mere 2 moves deep (your first move and your opponents first move) would give you approximately 500 moves to analyze. 2 more moves and you now have about 150,000 total moves. At 20 ply you have 1E26 number of moves, or a 1 followed by 26 zeroes.
Lastly, I sincerely do appreciate your posts and look forward to reading them. They are fun. Thanks!
|Jan-08-09|| ||patzer2: The necessary self-imposed pin and pin-busting maneuver 21...Qf7! (not 21...Qd7?? 22. Bxe6+ Kh8 23. Qd5! ) 22. Bxe6 Nxb3 adds slightly to the difficulty of the puzzle solution beginning with 18...d5!|
|Jan-08-09|| ||Lucid Faia: Got d5, but missed the follow-up.|
|Jan-08-09|| ||MostlyAverageJoe: <xrt999: I only run to 10 because I really dont grasp anything beyond that>|
The problem is that if you run engine to 10 plies, only the first 2 or 3 are of any use. When I run infinity analysis (btw, not on measly dual P4s, but on 8 Xeon cores :-) and get 24 plies deep, I would pay attention only to the first 6-8 plies at the most (except for some seriously forced lines).
The analysis which you did not like on another puzzle a while ago, was done primarily with forward/backward sliding, which is way more reliable than infinity analysis, and presented the optimal(*) line for both sides (the infinity analysis was done there only to get started). This was the kind of line and analysis that you could expect in correspondence chess nowadays.
(*) optimal, barring the situations where the real consequences are hiding behind the engine's horizon. Typically, forward/backward sliding will confirm winning lines, even if they elude the infinity analysis, but this is not always the case. For the best example where backslide fails to see the light, see Y Gusev vs Averbakh, 1951 where no engine I know of would approve a completely winning Q sac on move 24.
<what <I> consider real chess, OTB tourney play> Yes, this is one of realities, but CC chess is not less real, is it?
And yes, I can appreciate the reality of OTB, where you can expect that a line is good enough. In my last tournament, I played a line that was not even close to the best according to the post-game engine analysis, but got me into a R vs N endgame with many pawns which I expected to win. So yes, there are lines that are 'good enough' OTB, but they cannot be used to disprove an objectively best (with the same footnote * as above) line. Humans can sometimes find amazing moves (like this Gusev fellow) that may disprove a "good enough" move.
|Jan-08-09|| ||fouard: Shouldn't the goal be a clear win for Black? After the best analysis I have read here, which is 18...d5 19 exd5 Na5 20 d6 Qf7 21 Bxe6 Nxb3 22 Bxf7+ Rxf7 23 Rad1 Nd7, Black is only up a piece for 2 pawns, and unless he can win the passer soon, he may well be losing.|
|Jan-08-09|| ||johnlspouge: Thursday (Medium):
Anikaev vs A Khasin, 1990 (18…?)
Black to play and win.
Material: B for N. The White Kg1 has 3 legal moves. The Black Rc8 is on an open file; Rf8, on a semi-open file. The Black Nc6 and Nf6 are centralized, and Bg7 is on a semi-open long diagonal. Only the Black Qe7 requires activation. The Black Kg8 is vulnerable to the diagonal battery Qb3 and Qc4, which threatens the capture Bxe6+. Black can indulge in passive defense of Pe6, or he can counterattack.
Candidates (18…): Na5, d5
Because the capture by Bc4 is with check, I do not see how Black can gain a decisive advantage.
|Jan-08-09|| ||neilmcmurdo: This is a very tough puzzle. I'm surprised it didn't get a higher rating.|
Obviously it's easy to spot that the position comes down to 'does 18. ... d5 work?' But I think three aspects of the puzzle make it very tough:
- The simpler defensive moves for black are not obviously awful.
- To demonstrate that d5 works black has to spot the tricky Qf7 response to d6 in both the exd5 and Nxd5 lines.
- Moreover, having spotted that Black still has to conclude that the piece for two pawns exchange, with one of them advanced and passed, is actually good. And that at a depth of at least 6 moves in at least two different variations!
I've seen 'crazy' puzzles much easier than this one!
|Jan-08-09|| ||jovack: tricky puzzle|
|Jan-08-09|| ||MaxxLange: A miss for me today.|
|Jan-08-09|| ||e fred: What about:
18 ... d5
19 Bd3 d6
20 Bc4 Na5
21 Bxe6+ Kh8
22 Qa2 Rd8
or 19 Bg5
|Jan-08-09|| ||RandomVisitor: After 15...0-0
click for larger view
Analysis by Rybka 3 : 21-ply
1. (0.62): 16.Qd2 Rfd8 17.Rad1 Kh8 18.Rfe1 Na5 19.Bf1 Nc6 20.Bf2 Kg8 21.a3 Qf7 22.Nb5 Ne8 23.Rc1 a6
2. (0.55): 16.Rc1 Rfd8 17.Re1 Na5 18.Bf1 Nc6 19.Qb3 Qf7 20.Bg5 Rd7 21.Bb5 Rc8 22.Ne2 Rdc7 23.Bf4 Ne8 24.Rcd1
|Jan-09-09|| ||zenpharaohs: neilmcmurdo: "This is a very tough puzzle. I'm surprised it didn't get a higher rating."|
I think it's because the ratings are based on how long it takes a version of HIARCS to solve the position. In an example like this, HIARCS will compute the critical exchanges very quickly, which leads to an artificially low difficulty rating.
|Jan-09-09|| ||MostlyAverageJoe: <zenpharaohs: ratings are based on how long it takes a version of HIARCS to solve the position>|
NO, that's what <I> do when evaluating the puzzles. CG has a very different approach which they described on chessgames.com chessforum (and perhaps in other posts) and here: Daily Puzzle F.A.Q..
My evaluation places this puzzle on average Friday difficulty level.
< Earlier Kibitzing · PAGE 3 OF 3 ·