Originally posted by ivan2908Your sacrifice (21.Rxh6) was correct, but seems it would not give more than draw if Black would defend accurate.
This is a 3 minute RHP blitz. I thought my roc sac was sound. Everytime I am happy with my game, Fritz has to find 500 blunders in it. :'(
[pgn][Event "RHP Blitz rated"] [Site "www.redhotpawn.com"] [Date "2009.2.1"] [Round "?"] [White "ivan2908"] [Black "araykupo"] [Result "1-0"] 1. e2-e4 c7-c5 2. Ng1-f3 d7-d6 3. d2-d4 c5xd4 4. Nf3xd4 Ng8-f6 5. Nb1-c3 ...[text shortened]... g7xh6 22. Qd2-h2 Rc8-c7 23. Qh2xh6 Be7-f6 24. Bd4xf6 Rf8xf6 25. Rd1-h1 1-0[/pgn]
22...Rc7? was decisive mistake - better was 22....Bf6 and after 23.Qxh6 Qc7 24.g7! Bxg7 (24...Qxg7 25.Rg1 Bxd4 26.Rxg7 Bxg7 27.Nd5 and White should have at least perpetual check) 25.Bxg7 Qxg7 26.Qxe6+ Rf7 27.Nd5- despite of being rook down activity of White pieces + unsafe Black king will guarantee draw for White.
Position before your sacrifice seemed quite doubtful for you so your sacrifice was obviously the best move.
Something you rarely see on this forum - someone actually changing
their mind (me).
Though I still maintain constant playing them is harmful.
At least have a standard set next to the computer so you do not
lose your 3d vision.
But we won't go through all that again.
That ended Greenpawn 0 All the Geeks 5
I found more of my 'swindles' were infact sound and a few combo's
I thought were sound were infact 'dubious'. And that is using Fritz 6.
Eveluation may change when I upgrade.
I remember being given about 6 pirate copies of Fritz 10 or 11
to hand out to the boys in Bells.
I took them outside and one by one skimmed them onto the church
roof opposite Bells.
"These things do more harm than good." I proclaimed.
I wonder if they are still there?
Still stunned at Rybka finding 10.Ba6 and then evaluating it as sound
in me v R.Austin.
I played the thing because it looked trappy and interesting.
You spend you life wallowing in the myth you are a swindler and this,
this thing comes along and states it's a lie. Bah!
Perhaps if I wait 10 years. Fritz 30 will look at it and decide 10.Ba6
was unsound after all.
Originally posted by greenpawn34that's very true. even though we tend to think of engine evaluation as it was some kind of objective truth about the position, it isn't. just because engines tend to win the strongest humans now, that doesn't mean their evaluation is correct. not any more than a 1800s evaluation is absolutely correct just because he wins all games against 1200s. in the absence of a god-like absolutely correct reference evaluation, there's really no way to know whether an engine is any closer to perfection than a human. no matter how many GMs an engine beats, it can still be wrong.
Perhaps if I wait 10 years. Fritz 30 will look at it and decide 10.Ba6
was unsound after all.
there might very well exist far higher levels of chess than we can even think of now. something like 5000, 7000, 10000 (or arbitrary high) on FIDE scale, totally beyond anything we have now. we can't know. and if such higher levels turned up, it would mean that relative to that the contemporary engines make huge errors all the time. we just can't see it because we don't have a reference point.
the fundamental conclusion of this, until chess is solved, is that we can never know if an engine is correct or not. it might be statistically stronger than humans, but still a patzer relative to the 'truth of the position'. we just don't know, and can't find out.
Originally posted by wormwoodI didn't follow those discussions seriously, but in the Rybka forum, I remember discussions about God's elo (a philosophical god, playing the "best" move in all positions), and if I recall correctly, there was a consensus that its elo wouldn't be that high (like 5000). My uneducated guess would be something around 3500-4000 against human opposition.
that's very true. even though we tend to think of engine evaluation as it was some kind of objective truth about the position, it isn't. just because engines tend to win the strongest humans now, that doesn't mean their evaluation is correct. not any more than a 1800s evaluation is absolutely correct just because he wins all games against 1200s. in the abse tzer relative to the 'truth of the position'. we just don't know, and can't find out.
I know I seem like a fanboy, but in the opening with no book, Rybka 3 is getting VERY consistent results with hundreds of years of human theory when it's unaided (no books), and I believe that indicates it's seriously close (several decades) to perfrection in the middlegame. Endgames though might still have a long way to go in terms of raw engine evaluation, but probably tablebases would compensate in that area.
I don't think you are a fan boy - you spun me around.
(of course you will get the geek jokes - but I get the dinosaur jokes).
It's an intresting point about God's grade.
It may be quite low. 1600ish as he has a lot more important things to
do than study chess.
I'm of the opinion that computers will bust chess eventually but most
people will just shrug and carry on playing anyway.
(I hope they don't switch to Random Chess - I hate it)
Originally posted by wormwoodI think opening theory could serve as a good reference point for comparison, I know there's no necessity or scientific laws yet, but it seems to me that a large portion of theory is close to perfection. but defining perfection is difficult too, so I'll keep it at "being close" only. but negative comparisons could very well work.
the fundamental conclusion of this, until chess is solved, is that we can never know if an engine is correct or not. it might be statistically stronger than humans, but still a patzer relative to the 'truth of the position'. we just don't know, and can't find out.
for example, if after 1.e4 e5 2.Nf3 Nc3 3.Bc4 Nd4?, if an engine grabs the pawn with 4.Nxe5 ??, although we don't yet have all the lines played out (i.e "solved" ) and although this move and all the other moves could reach a draw with "perfect play", we can assume that engine is far, far away. (there should be more subtle and better examples, but I know very little theory.)
I appreciate the point wormwood is trying to make - there are categories of chess strength unknown to us today. However I still would place the theoretical upper limit of chess strength - "God's rating" to use a euphemism - to still be fairly low, say around 3200-3300. The possibility of draws is inherent in what chess is - and the theoretical monster machines of 20 years from now wouldn't be able to win every single game against human+computer centaurs playing today - not if they're aiming to make a draw.
It's possible that I'm not picking the precise endpoint. My main point is that chess strength is asymptotically approaching perfection, and the point is perfection isn't leaps and bounds above where it is today. What would a rating list for Tic-Tac-Toe look like? And draughts has extraordinary depth, but it's now a solved game and all positions can be evalutated perfectly.
Now, expand the board to 12x12 and add four new pieces, and then we'll see how much better the future machines are than what's capable today.
Originally posted by greenpawn34I agree that when it gets solved, it for sure will be a negative thing. But I doubt it would change how I study or play. I'm trying to get decent in the english for 2 years, and I'm not even close.
I'm of the opinion that computers will bust chess eventually but most
people will just shrug and carry on playing anyway.
(I hope they don't switch to Random Chess - I hate it)
I don't like FRC either, but I don't think we have a better choice. either a variant has to be agreed upon globally (which seems very unlikely), or the starting position has to change. These are, of course, after things like changing the castling rule, adding forced promotion to queen, or removing en pessent rule are tried to keep the game alive longer.
Originally posted by DawgHaus3200 is much too low. Rybka with a pawn or the exchange down plays around 2700 level right now. If she would go in the FIDE pool right now, I think 3100 would only be a matter of time if she could choose her tournaments (like super GMs do). We can assume Rybka 5 or Fritz 14 to be 3200.
I appreciate the point wormwood is trying to make - there are categories of chess strength unknown to us today. However I still would place the theoretical upper limit of chess strength - "God's rating" to use a euphemism - to still be fairly low, say around 3200-3300. The possibility of draws is inherent in what chess is - and the theoretical monster mac and then we'll see how much better the future machines are than what's capable today.
Originally posted by diskamylLet Black go first. .😉
I agree that when it gets solved, it for sure will be a negative thing. But I doubt it would change how I study or play. I'm trying to get decent in the english for 2 years, and I'm not even close.
I don't like FRC either, but I don't think we have a better choice. either a variant has to be agreed upon globally (which seems very unlikely), or the start ...[text shortened]... ced promotion to queen, or removing en pessent rule are tried to keep the game alive longer.
Serioulsy - swap the Black King and Queen about. King d8 - Queen on e8.
I have a game played in a low graded tournament where Black
did this and it went unnoticed by both players.
True - it's on the Corner somewhere.
The game, as I remember look quite 'normal' Black 0-0 (which was
really 0-0-0) and got mated with a standard Queen and Bishop battery.