< Earlier Kibitzing · PAGE 4 OF 4 ·
|Jan-04-08|| ||A.G. Argent: <UdayanOwen; Have to go into early retirement>; While you may need to step away from this site and distance yourself from the analysis jones and deal with such a minor element in your life like a new baby, I heartily doubt you ain't comin' back at all, judging from some previous posts. You're the one who used the word addiction. Use the word hiatus instead of retirement. Besides, if you're reading this, you ARE back. And since you are back, who was it you quoted ?: "Chess is not a game. It is a profound and beautiful universe....." Anybody else know who said it?|
|Jan-04-08|| ||Jimfromprovidence: 29 Rc3! is the best move nobody has ever played. Ain’t it a shame that a computer figured it out?|
<Maj> <After 11 CPU hours, Hiarcs insists that
1. (+5.10) 29. Rc3 Rc8 30. Rxg6 ...
is the best continuation,>
I’m having difficulty finding the way to a rook’s worth gain of material in this line.
What’s the continuation after 30… hxg6 31 Qxg6+ Kf8 32 Rg3 Bxg2?
|Jan-04-08|| ||znprdx: <parmetd: wow I got this so wrong...> not at all in fact you got it so right <None of the moves played seemed like a forced variation.>|
d5 was an obvious ploy to get the rook to c7. The basic (yawn) Rc3 rook lift equally straight -forward - which I'd play in tournament..but end up struggling because of complications. >The only real finesse here would be 29.Bd6<which I'd play in Rapids to keep the King off f8, then after the forced win combo of Rxg6 almost anything wins.
Also what was to stop Black from trying a desperation strategy after the dubious 30.Bb2? with ...Qxg2+ ...I think this at least draws...In Blitz I'd have probably played 29.Rxg6 immediately
Happy New Year <MAJ> - if there is a limit to the usefulness of computer analysis - we can see its shortcomings here - sometimes there are just too many possible reasonable continuations to insist on any absolute verdict, which in this case is that position was pretty one-sided - and <CG>'s choice does NOT really meet a nominal criteria of being a puzzle -
|Jan-04-08|| ||futonchild: Shouldn't that read "36. Qe7#"?|
|Jan-04-08|| ||alphee: Very nice puzzle. It made me search for a while and I missed d5. I should have seen it as I was looking for a way to deviate the Queen to give to the suite ♖xg6+, hxg6, ♕xg6+, ♔f8 a chance to work.|
|Jan-04-08|| ||UdayanOwen: <A.G. Argent: <UdayanOwen; Have to go into early retirement>; While you may need to step away from this site and distance yourself from the analysis jones and deal with such a minor element in your life like a new baby, I heartily doubt you ain't comin' back at all, judging from some previous posts. You're the one who used the word addiction. Use the word hiatus instead of retirement. Besides, if you're reading this, you ARE back. And since you are back, who was it you quoted ?: "Chess is not a game. It is a profound and beautiful universe....." Anybody else know who said it?>|
I will pop in having a look at puzzles and saying hi, but if I start kibitzing the taste will bite me and I'll get drawn helplessly into hours of analytical work, sacrificing sleep, neglecting my wife, children, chores and university studies.
I'm not giving up chess, I'll just do studies that I can put down when I need to.... chessgames kibitzing doesn't fit into that category for me.
That phrase you alluded to came from my head... if someone else said it, that is probably a coincidence (alternatively, I may have read it without remembering it, and it came through subconsciously.... I'd probably remember it though, so I doubt this explanation).
I prepared several other, even more over the top 'poetic' odes to the beauty of chess.... but held back on the posting when I concluded they were totally corny and lame....
P.S. Chess STILL @#$%ing rocks!!!!
|Jan-04-08|| ||wals: Alexander Baburin - Sergey Fokin, URS-sf RUS-ch Gorky 16 1989
click for larger view
Analysis by Fritz 11:
1. (2.24): 29.Rc1-c3 Ra8-c8 30.Rg4xg6+ h7xg6 31.Qg3xg6+ Kg8-f8 32.Rc3-g3 Bc6xg2 33.Rg3xg2 Qd7-d8 34.Qg6-h5 Qd8-d7 35.Be5-g7+ Kf8-e7 36.h6-h7 Ke7-d6 37.h7-h8Q Rc8xh8 38.Bg7xh8 a7-a5 39.Bh8-e5+ Kd6-c6 40.Rg2-g3
2. (1.92): 29.Be5-d6 Ra8-c8 30.Rg4xg6+ h7xg6 31.Qg3xg6+ Rf7-g7 32.h6xg7 Qd7xd6 33.Rc1-c3 Qd6-d5 34.Rc3-f3 Qd5xf3 35.g2xf3 Bc6-d5 36.Qg6-f6 a7-a5 37.Kg1-g2 a5-a4 38.Qf6-e7 b6-b5 39.Kg2-g3
|Jan-04-08|| ||MostlyAverageJoe: <Jimfromprovidence:
<Maj> <After 11 CPU hours, Hiarcs insists that 1. (+5.10) 29. Rc3 Rc8 30. Rxg6 ...
is the best continuation>
What’s the continuation after 30… hxg6 31 Qxg6+ Kf8 32 Rg3 Bxg2>
Here's a quick forward scan:
<29. Rc3 Rc8 30. Rxg6+ hxg6 31. Qxg6+ Kf8 32. Rg3 Bxg2> 33. Rxg2 Qd8 34. Rg4 Rc1+ 35. Kh2 Rxf2+ 36. Kg3 Rcf1 37. Qg8+ Ke7 38. Rg7+ Rf7 39. h7 Rf3+ 40. Kh4 Qd5 41. h8=Q
click for larger view
This was done very quickly, so improvements are possible. The line shows white's primary threats of Qg8 and h7-h8=Q. If you play through it, it is quite amazing how a naked white K escapes unscathed. This line is really for computers only :-)
41. ... Qe4+ 42. Rg4 Qe1+ 43. Kg5 Qd2+ 44. Bf4 Qa5+ 45. Qe5
and black is out of useful checks. Hiarcs declares mate in 14 from here.
White has a bit more trouble after B sac in move 32:
32. ... Be4 33. Qxe4 Qc6 34. h7 Qxe4 35. h8=Q+ Ke7 36. Qxc8 Qb1+ 37. Kh2 Rh7+ 38. Rh3 Rxh3+ 39. Kxh3 Qh7+ 40. Kg3 Qg6+
but the material gain is clear here:
click for larger view
|Jan-04-08|| ||MostlyAverageJoe: <znprdx: Happy New Year <MAJ>>|
And Happy New Year to you too!
Too bad I have to disagree right away with <if there is a limit to the usefulness of computer analysis - we can see its shortcomings here> -- IMHO, this puzzle shows exactly how computers can outmaneuver (most) humans in tactical positions.
MAJ, thinking about partial retirement from CG for similar reasons to <UdayanOwen>'s -- this site is too @#$%ing addictive.
|Jan-04-08|| ||Jimfromprovidence: Thanks to <wals> and <maj> for their positional analysis of the line I put forward. <<What’s the continuation after 30… hxg6 31 Qxg6+Kf8 32 Rg3 Bxg2?>> |
They do digress on move 34 for white, though. Don't know what that means, however).
I love the rook lift but who has the guts and knowledge to play it correctly?
Your opponent knows exactly what you’re doing so you have to be pretty certain that it's going to work before you try to pull it off.
|Jan-04-08|| ||dzechiel: <parmetd: dzechiel post indicates the draw is guaranteed simply cause of the opposite colored bishops.>|
What I was trying to convey (obviously not very successfully) is that we white shouldn't go down a line that trades the major pieces off the board, as a bishop and pawn endgame would be difficult to win.
All of this was meant to imply that white should be looking for moves that either win an exchange (no more opposite colored bishops) or go right for the jugular with a mating attack.
|Jan-04-08|| ||Gouki: <Shouldn't that read "36. Qe7#"?>|
Yes it should. The position at the end is mate.
|Jan-04-08|| ||MostlyAverageJoe: ,Jimfromprovidence: Thanks to <wals> and <maj> for their positional analysis of the line I put forward. <<What’s the continuation after 30… hxg6 31 Qxg6+Kf8 32 Rg3 Bxg2?>>
They do digress on move 34 for white, though. Don't know what that means, however).>|
The Fritz analysis from <wals> is an infinity analysis starting at move 29 -- and I am pretty sure that it is less than 30 plies deep. Therefore, move 34 is definitely not to be trusted there.
My line is done in forward scan mode, where I run in mullti-PV for a while and follow the best choice. This is equivalent to putting the engine on auto-play with fixed search depth per move.
Forward scan accelerates the analysis at least an order of magnitude, but possibly misses some better choices due to the shallower analysis (e.g., 16 plies) in each move. Sometimes the moves are forced, though, and this is where such scan can be faster than auto-play. Lines obtained this way are good candidates for backsliding (to identify missed opportunities).
Some backsliding on my initial line is what indicated that 32...Bg2 is sub-optimal, too, and 32...Be4 sac appeared to be better.
|Jan-04-08|| ||DarthStapler: I kept looking at Rxc6 to deflect the queen from the 7th rank|
|Jan-04-08|| ||johnlspouge: Just to add more computer confusion into the mix, I left Toga II 1.3.1 running during the day. |
18 ply 0:54:21 time +4.38 value
29.Rc3 Rc8 30.Rxg6+ hxg6 31.Qxg6+ Kf8 32.Rg3 Bxg2 33.Rxg2 Qd8 34.Rg4 Rc1+ 35.Kh2 Rxf2+ 36.Kg3 <Rg2+> 37.Kxg2 Qd5+ 38.Kg3 Rg1+ 39.Kf2 Rxg4 40.Qxg4 Qa2+ 41.Kf3 Qa3+ 42.Ke4 Qa2 43.Bd6+ Ke8 44.Qg8+ Kd7 45.h7 Qe2+ 46.Kf4 Qf2+ 47.Kg5 Qf5+ 48.Kh4 Qf2+ 49.Kh3
The Black move 36...<Rg2+> diverges from <MAJ>'s Hiarcs analysis.
Interestingly, later Toga changes its mind on White move 30.
21 ply 3:39:46 time +4.23 value
29.Rc3 Kf8 30.<Rf4> Rxf4 31.Qxf4+ Ke8 32.Bd6 Qf7 33.Qc1 Rc8 34.d5 exd5 35.Rxc6 Rxc6 36.Qxc6+ Qd7 37.Qxd5 Qf5 38.Qa8+ Kd7 39.Bb8 g5 40.Qb7+ Ke8 41.Bxa7 Qb1+ 42.Kh2 Qg6 43.Bxb6 Qd6+ 44.g3 Qxh6+ 45.Kg2
Just in case anyone wants to question the competitiveness of freeware, Toga II 1.3.1 ($0, free) took the toughest of puzzle I have seen
G M Todorovic vs Z Kozul, 1988
and came up with the correct Rh5 Rxd5 maneuver at 17 ply, 1:25:32 time, +6.02 value. Rybka 2.3.2 ($40) still had not got the maneuver after 25 ply 9:46:55 time +3.49 value.
|Jan-04-08|| ||DevastatioN: <Chessgames.com> It should be Qe7# not Qe7+ :)|
|Jan-04-08|| ||A.G. Argent: <UdayO> Now obviously I liked your adage/observation/epigram whatever. So as not to belabor the matter further in this forum, could you maybe drop a few of your others in my forum. Doubt I'd find 'em corny or over the top. I love that stuff. Thanks pal.|
|Jan-04-08|| ||johnlspouge: <patzer2: an even stronger and more human-like solution is 29. Rc3!!>|
I suspect more of us would have looked at 29.Rc3 if it had been Sunday ;O}
|Jan-04-08|| ||whiteshark: <29.Rc3> is rather impressive. It's a good example for attacking when you have opposite coloured bishops aside heavy pieces. Here you have the additional power of a ♙h6.|
After <29.Rc3 Rc8 30.Rxg6+ hxg6 31.Qxg6+ Kf8 32.Rg3> <32...Be4> looks only like the lesser evil in a hopeless position.
|Jan-04-08|| ||johnlspouge: Interesting. The computer rates B vs Ps inferior, but Q+B vs Q+Ps superior. There is something (obvious, in retrospect) to be learned. A piece vs. pawns might generally be better in the middle game than in the endgame, because of mate threats.|
|Jan-04-08|| ||zooter: Spike 1.1 has the following analysis in favor of 29.Rc3|
+4.24 29.Rc3 Kf8 30.Rf4 Rd8 31.d5 Qxd5 32.Rd4 Qxd4 33.Bxd4 Rxd4 34.Rxc6 Rd1+ 35.Kh2 Rdd7 36.Rxe6 Rfe7 37.Rf6+ Rf7 38.Qb8+ Ke7 39.Qe5+ Kf8 40.Rc6 Rd8
What was interesting is that after about 9 hours of analysis, the ply is only at depth 21 with about 28 Billion moves analysed! This I feel shows the sheer complexity of the position as somebody else pointed out
|Jan-04-08|| ||znprdx: <MAJ> I'm kind of dismayed you still don't get it <11 CPU hours> - how many days - months - or years is that in human time? |
It is increasingly pathetic to watch people posting computer analysis while watching the world's top players in a tournament or match. As I've posted more than once, save it for the final draw positions - and/or maybe even incorporate it - into deciding their outcomes.
Other than that there is relatively little value - particularly in going anywhere beyond 12 ply in maybe 6 variations ...because humans just don't do that: they intuit the implicit synergy. They move and in a finite period of time the opponent must move - and that's all folks. Either we incorporate tracking all the analysis a player evaluated - or we leave our beautiful game uncontaminated....
|Jan-04-08|| ||GannonKnight: I was also zeroed in on Rxc6 as a deflection. But there are better ways to get the Black queen out of the way and then to set up Ba3+. Nice, but terribly difficult puzzle.|
|Jan-04-08|| ||RandomVisitor: Black's last chance to equalize was apparently 27...Rc8, which stops 28.Rc1.|
|Jan-05-08|| ||MostlyAverageJoe: <znprdx: <MAJ> I'm kind of dismayed you still don't get it <11 CPU hours> - how many days - months - or years is that in human time?>|
Hmm, I guess you did not look at my first post, which reported 29.Rc3 as the best move after 16 plies. Actually, it got ahead of the alternate moves after 15 plies, or about 90 seconds of analysis; at 16 plies, after 3 minutes of running, the move appeared to be a clear winner.
The 11 CPU hours was just because I had to sleep instead of driving the engine. With some manual nudging, I'd get to the same conclusion much faster, but I did not feel like spending time on that, while wanting to see the results of deep analysis, especially after noticing a rather slow start (only 15 plies in 90 seconds is unusual in positions with forcing tactics).
It is noteworthy that the results did NOT change - the same move picked after 90 seconds stayed on top; it was only the subsequent line and the final valuation that got improved.
The remaining analysis I posted took me maybe 15 minutes to do, and even if it contained some inaccuracies, I bet they were quite small.
Thus, your dismay over 11 hours is not justified. I'd venture a guess that at human time controls, the engine would easily have found the best move, particularly since it would not have to start from scratch, but would have the benefit of knowing the previous moves and analysis of the possible continuations.
As a matter of fact, I just run a quick experiment, starting two moves earlier, and following the game at 30 seconds per move. Once it came to move 29, the 29.Rc3 was picked just about instantaneously.
<It is increasingly pathetic to watch people posting computer analysis while watching the world's top players in a tournament or match>
This is a completely different topic. By the way, it is probably easier to find a chess player whose intuition will result in picking the winning lines, than to find a human capable of writing software that will find the same lines. But then again, as a software wrangler myself, I have a different appreciation of that domain than many others.
< Earlier Kibitzing · PAGE 4 OF 4 ·