|
< Earlier Kibitzing · PAGE 2 OF 3 ·
Later Kibitzing> |
| Jan-13-08 | | SpoiltVictorianChild: Can someone tell me: Is 25 Nd5 a mistake? If he had gone straight for the pawn, could he have kept both rooks? Even if possibly losing the knight? |
|
| Jan-13-08 | | Aurora: My winning line is <18.Qe3 cxd3 19.fxg6 hxg6 20.Qh6 Bg5 21.Qxg5 Nc4 22.Nd5 exd5 23.Bd4 Ne3 24.Qxe3  > |
|
| Jan-13-08 | | znprdx: Oh yes now we are talking -this is a truly extraordinary combination. I ran out of steam at 7ply at which point the real puzzle begins. 21.Qg4!!!!!!!!!!! now this was insane....and very rare.I almost had to set it up to believe it...21. e6xd5[N]22. Rxf6 threatening everything, which is why e5 was played (to no avail)...22.Bxg4[Q} 23. Rf7+!!!! at 11 ply leads to mate.<CG> has been redeemed after some pretty lame non-problems of late.... |
|
| Jan-13-08 | | johnlspouge: Sunday (Insane) White to play and win
Material even. White's Bs rake long diagonals, particularly Bb2 highlighting the weak dark squares around Black's K, the Nc3 can reach weak dark squares via e4 or d5, and the Qe2 and Rc1 are poised for K-side attack, spearheaded by the Pf5. White's Bd3 is under attack from Pc4, and must move [18.bxc3 Rxb2], and the Pf5 is 4x attacked-2x defended, indicating that immediate action is required or the spatial advantage will dissipate. Candidates: Nd5 (N-leap week at CG.com?), f6, fxg6, fxe6 The big question is how to reply to cxd3, if White does not play the vanilla Be4, which loses? For example, 18.Nd5 cxd3 19.Nxe7+ Qxe7 20.Qe5 f6
gives nothing. This, I imagine, is a rudimentary example of the process <UdayanOwen> called "sampling a line". Only one candidate creates an threat immediate enough to let the Bd3 hang, and that is 18.fxg6 cxd3 19.gxh7+ Kxh7 [Kh8 or Kg7 20.Nd5+ throws the Q for 2Bs] 20.Qh5+ Kg8 [else 21.Nd5+ again] 21.Nd5 toasts Black [lose the Q or Qg7#] Thus, 18...cxd3 is impossible, and either the f- or h-files become available for White's Q and R. This line is enough support to justify making the move fxg6, as White's position improves [30 min analysis]. Let's go a little further. The move 20...f6 that refuted 18.Nd5 becomes impossible on 18...fxg6. 18.fxg6 fxg6 19.Nd5 cxd3 [exd5 20.gxh7#] 20.Nxe7+ Qxe7 21.Qe5 and White will # with Qg7 or Qh8. If 19...Q move, 20.Nxe7+ goes down the same line. Thus, the only (possibly) viable line for Black is
18.fxg6 hxg6
There is no obvious forcing continuation, so <dzechiel>'s iterative method of selecting candidates comes into play. Candidates: Nd5, Qh5, Bxg6
The obvious candidate is Bxg6, to eliminate the ...f6 defense again. 19.Bxg6 fxg6 20.Nd5 exd5 21.Qe5 Bf6 22.Rxf6 Rxf6 23.Qxf6 and White wins because of the threat of Rf1 and then # [60 minutes analysis]. Time to peek. I missed 20...f6 as a viable defense. Time to review the kibitzing. I missed 20...f6, so I never foresaw the brilliant 21.Qg4. Truly beautiful!!! Obviously, Van der Wiel calculates through variations faster than I do, but maybe in a few more Sundays, with a little more practice... ;>) |
|
| Jan-13-08 | | znprdx: <dzechiel: I don't understand why 21...e5 instead of 21...exd5. What am I missing here?> Hi (Happy New Year!!!) see my post. I'm not convinced this was calculated from 18. so much as it was intuited - it is almost mesmerizing. Seeing the idea of an eventual Rf7 is phenomenal. |
|
| Jan-13-08 | | johnlspouge: <Manic: What is are the variations after 18...fxg6 and 18...hxg6?> Toga II 1.3.1 analyses
[ply 15, time 3:49, value -5.68]
18...fxg6 19.Qe5 Rf6 20.Ne4 Ne8 21.Nxf6+ Bxf6 22.Qxc7 Nxc7 23.Rxf6 Nd5 24.Bxc4 Nxf6 25.Bxf6 Rb6 26.Be5 Bd7 27.Rd1 Bb5 28.Bc7 Rc6 29.Rd8+ Kf7 30.Rd7+ Kf6 [ply 15, time 8:15, value -3.03]
18...hxg6 19.Bxg6 fxg6 20.Qe5 Rxf1+ 21.Rxf1 Ne8 22.Nd5 Qxe5 23.Nxe7+ Kh7 24.Bxe5 Rb5 25.Bb2 c3 26.Bxc3 Rc5 27.Rf7+ Kh6 28.Bd2+ g5 29.c4 Nd6 30.Rf6+ Kg7 31.Rg6+ Kf7 32.Rxg5 The times taken for the 15-ply analyses are extraordinary, presumably reflecting the unusual complexity of the position. In case you want to do your own computer analyses, my chessforum johnlspouge chessforum
gives explicit instructions on how to download freeware to do chess analysis. I recommend the download: nowadays, unless demonstrating human chessplaying processes, most human analysis is worthless compared to computer output. |
|
Jan-13-08
 | | An Englishman: Good Evening: I actually got this one, which is bizarre. |
|
| Jan-13-08 | | johnlspouge: <znprdx: I'm not convinced this was calculated from 18. so much as it was intuited - it is almost mesmerizing.> I am almost sure that where necessary, variations were calculated to a winning conclusion. I am only starting to train myself to visualize deep plies, but the best humans must be able to calculate the variations here to finality within 30 minutes. Today's position had an unusual characteristic, in that one kept seeing the thematic candidate moves recurring in different variations (e.g., Nd5, Qe5, Nxe7). The recurrence of the candidates speeds exact calculation, because the moves become familiar elements in solving a problem. |
|
| Jan-13-08 | | johnlspouge: <stukkenjager: 19.Qh5 looks ok to me> Toga II 1.3.1 ply 15, time 4:36, value +2.26
19...hxg6 20.Qh6 Bg5 21.Qxg5 Nc4 22.Nd5 exd5 23.Bd4 Ne3 24.Qxe3 dxc2 25.Qh6 f6 26.Qxg6+ Qg7 27.Qxc2 Bd7 28.Rad1 Qg5 29.Rde1 Kh8 As a baseline, from the puzzle position
Toga II 1.3.1 ply 15, time 9:55, value +1.77
18.fxg6 cxd3 19.gxh7+ Kh8 20.Nd5+ f6 21.Qg4 e5 22.Qg8+ Rxg8 23.hxg8Q+ Kxg8 24.Nxc7 Bd8 25.Nd5 d2 26.Rad1 Bg4 27.Rxd2 Ne4 28.Rd3 Be2 29.Re3 Bxf1 30.Rxe4 The your line's value looks comparable to the game line's, which the computer follows from the puzzle position to 24...Bd7 [24...d2 in the game]. The different values might be due to the different starting move numbers with fixed ply, of course. |
|
| Jan-13-08 | | midwood2001: what about 26.Nxf6+?
click for larger view |
|
| Jan-13-08 | | johnlspouge: <Aurora: My winning line is 18.Qe3 cxd3 19.fxg6 hxg6 20.Qh6 Bg5 21.Qxg5 Nc4 22.Nd5 exd5 23.Bd4 Ne3 24.Qxe3> 18.Qe3 leaves White +0.52:
Toga II 1.31. [plies 15, time 20:31, value +0.52]
18.Qe3 cxd3 19.fxg6 f5 20.Qh6 Bd8 21.Ba3 Rb7 22.Ne4 Qg7 23.Qxg7+ Rxg7 24.Nxd6 Bf6 25.gxh7+ Rxh7 26.Rac1 Bg5 27.cxd3 Bxc1 28.Rxc1 Bd7 29.Bb2 Computers do not give perfect evaluations, of course, but the point of my rash of computer posts is that you can do more than speculate on your solution's value. Consult the oracle ;>) |
|
| Jan-13-08 | | MaczynskiPratten: <midwood2001: what about 26.Nxf6+?> I wondered this too. I think the answer is Bxf6 27 Rxf6 Ne4 and White has a remarkably hard time finding a safe place for his Rook. Probably best is Rf1 but then Bg4 will win back the exchange in a similar way to the game - if Black wants to. His Ne4 and Pd2 are so powerful that he might want to keep them and tie White up. Hence White's decision to eliminate the d2 pawn at once and give back the exchange. It would be interesting to turn the silicon monsters loose on these lines and see what they think was best. Also, the rest of the game is interesting. White's extra 2 pawns do not seem to make a techncal victory a foregone conclusion in a rook endgame; in fact he exploits his superior mobility to win tactically. |
|
| Jan-13-08 | | johnlspouge: <MaczynskiPratten>: OK, but this is it for today. [ply 15, time 0:32, value +1.40]
26.Nf6+ Bxf6 27.Rxf6 Ne4 28.Rh6 Nf2+ 29.Kg1 d1Q+ 30.Rxd1 Nxd1 31.Bxe5 Rb5 32.Bd4 Rd5 33.c3 Kf7 34.Rc6 Rd8 35.h4 Rg8 36.Rf6+ Ke7 37.Rf1 Bg4 38.c4 Kd6 39.Rf7 Re8 [ply 15, time 0:17, value +2.09]
26.Rad1 Bg4 27.Rxd2 Ne4 28.Rd3 Be2 29.Rdf3 Bxf1 30.Rxf1 Rb7 31.c4 a5 32.Bc1 Kf7 33.Be3 Be7 34.g4 a4 35.Bb6 axb3 36.axb3 Bb4 |
|
| Jan-13-08 | | Aurora: <johnlspouge> Thank you for looking! <19...f5> is a real change for the better. Could you ask your oracle again on my new line <18.Qe3 cxd3 19.fxg6 f5 20.Qh6 Bd8 <21. Rf3 Rf6 22.gxh7+ Qxh7 23.Rg3+ Kh8 24.Qf4>> ? |
|
Jan-13-08
 | | Jimfromprovidence: An absolute superb puzzle. One of the best I've seen on this site. Speaking of move 26 for white, I actually like 26 Rxf6! because it wins a third pawn for white. If 26...Bxf6 then 27 Nxf6+ Kf7 28 Bxe5. 
click for larger view |
|
| Jan-13-08 | | Anatoly21: Well, I got it, though certainly not all the way out. On all the Insane puzzles on this site I've given up calculating it out and go with my intuition, and in this instance got me out to 21. Qg4, where I thought white must be winning somewhere. Its positions like these where OTB I, and probably most individuals here, would have just said "This must be right" and gone with it. |
|
| Jan-13-08 | | johnlspouge: <Aurora: <johnlspouge>...Could you ask your oracle...?> The oracle accepts only one request per customer per day (if you are very lucky)! I am encouraging people to go to my chessforum
johnlspouge chessforum
which has extensive instructions on how I downloaded freeware for chess analysis. You can also go to User: crafty and ask chessgames.com for specific analyses from the computer program "crafty". (chessgames.com are making money from this site; I am not ;>) |
|
| Jan-13-08 | | InspiredByMorphy: <johnlspouge> <18...hxg6 19.Bxg6 fxg6 20.Qe5 Rxf1+ 21.Rxf1 Ne8 22.Nd5 Qxe5 23.Nxe7+ Kh7 24.Bxe5 Rb5 25.Bb2 c3> If instead black plays 25. ...cxb3 everything changes. 
click for larger viewNow if 26.Nxc8 bxc2 27.Bc1 Rd5 black is still in the game. Black's passed pawn probably being at least as equal in value as whites extra piece. 
click for larger viewAfter 25. ...cxb3 White is better off playing 26.cxb3 where one plan of action for black could be 26. ...Bb7 27.Rf7+ Kh6 28.Rf8 Rc5 
click for larger viewwith threats of ...Rc2 and ...Rxg2 black appears to still be in the game. All in all it appears that 18. ...hxg6 was the correct reply for black. |
|
| Jan-13-08 | | InspiredByMorphy: <johnspoulge> Perhaps in the line above white is better off playing 26.axb3
where one continuation may be ...Rc5 26.c4 e5 where keeping blacks rook out of the action may help maintain whites positional + 1 pawn advantage.
click for larger view |
|
| Jan-13-08 | | johnlspouge: <InspiredByMorphy: <johnspouge>> Thanks for your analysis.
I included the computer line to the end according to the principle of not destroying information when communicating (scientific) results. Of course, near the end of its plies, the computer line is going to miss much that a human can see. In an exceptionally complicated situation like today's, it would be desirable to run computer analysis for hours rather than minutes, to check that the evaluation is stable as the plies increase. I recommend that a few other people download freeware, so we can devote a few more computer cycles to the analyses ;>) |
|
| Jan-13-08 | | johnlspouge: <InspiredByMorphy: All in all it appears that 18. ...hxg6 was the correct reply for black.> I ran Toga II 1.3.1 for an extra ply
[ply 16, time 13:40, value +3.10]
18...hxg6 19.Bxg6 fxg6 20.Qe5 <Ne8> 21.Nd5 Qxe5 22.Nxe7+ Kh7 23.Bxe5 Rxf1+ 24.Rxf1 Rb5 25.Bb2 cxb3 26.axb3 Rc5 27.Rf8 e5 28.Rxe8 Rxc2 29.Nxc8 Rxb2 30.h4 Rxb3 31.Re7+ Kh6 32.Rxe5 Interestingly, the highlighted 20...Ne8 diverged from earlier analysis (and was present at ply 15, time 0:35, value +2.98, as well). I have not discovered any detailed description of the Toga algorithm, but the program appears to learn, because it found its 15-ply line much faster than before. I will let the program run some more, and let you know if there is a drastic change in the evaluation. Personally (having done a little human analysis on the position as well ;>), the variation played out to 21.Qg4 was probably best play. It is hard for me to believe that Black is better off with 18...hxg6 than 18...cxd3. With 18...hxg6, he lets the Bd3 sacrifice itself for a valuable defensive K-side P, and without 18...cxd3, the Pc4 never becomes a dangerous passer. |
|
| Jan-13-08 | | johnlspouge: <InspiredByMorphy>, Toga II 1.3.1 [ply 18, time 1:27:36, value +3.32] gave the same moves as ply 15 out to 26...Rc5. I know by experience that each ply doubles to triples the previous calculation time, and I would like my computer back ;>) |
|
| Jan-13-08 | | MaczynskiPratten: I think over the board 18..cxd3 is better than 18...hxg6 because it is easy for White to follow up after the latter, but with the former he has to find the very difficult 21 Qg4! to keep the attack going effectively. |
|
Jan-13-08
 | | Cinco: <johnlspouge>, Fritz likes two lines here. 18.f5xg6 c4xd3 19.g6xh7+ Kg8-h8 20.Nc3-d5+ f7-f6 21.Qe2-g4 e6-e5 22.Qg4-g8+ Rf8xg8 23.h7xg8Q+ Kh8xg8
± (1.22) Depth: 18/42 230mN
And
18.f5xg6 c4xd3 19.Qe2-h5 h7xg6 20.Qh5-h6 Be7-g5 21.Qh6xg5 Nd6-c4 22.Nc3-d5 e6xd5 23.Bb2-d4 Nc4-e3 24.Qg5xe3 f7-f6 25.Rf1xf6 Rf8xf6 26.Bd4xf6 Bc8-f5
(1.97) Depth: 18/42 23796kN
Thoughts on this second line with 19. Qh5? |
|
| Jan-13-08 | | johnlspouge: Hi, <Cinco>. Welcome to chessgames.com! Since you have your own oracle, I assume you are asking me for a human evaluation :) My analysis did not consider 19.Qh5 as a candidate against 18...cxd3, but this was an oversight on my part. The counter 20...f6 is what necessitated the second key move 21.Qg5 in the game. Over the board, therefore, if you do not have Van der Wiel's calculating abilities (or his intuition, maybe), 19.Qh5 would be a powerful alternative to 19.gxh7+, as it permanently excludes ...f6 (because of Qxg6+), thereby simplifying the lines enormously. As I calculated, I was impressed that certain thematic moves kept recurring in different variations. In practice, I would have excluded ...f6 gladly, even if it cost a few centi-Ps :) As far as I know, Toga does not offer alternatives to its best lines, so thank you for taking the hint and providing your own computer analysis. |
|
 |
|
|
< Earlier Kibitzing · PAGE 2 OF 3 ·
Later Kibitzing> |
Kibitzing
sac: 18.fxg6