bob wrote:
Feel free to give some actual evidence to support such a claim. AGAIN, not even Larry claims that, except that for programs/hardware today a knight is too much. Barely.
he said
" Of course I'm just giving my opinion on how fast the series will converge; it is of course possible that it will reach 3.5 (knight odds) some day. But my feeling is that Komodo at tournament level is roughly one pawn away from perfect play, which would mean that we might eventually be able to give Carlsen two pawns or a bit more, but not a knight. There are tests we could run to try to confirm this, but it would detract too much from Komodo to do this"
bob wrote:In the case of the knight, if he trades perfectly, at best he can draw, since using your rather ridiculous analysis, KNK is a dead draw.
So you still have not grasped the rather simple argument. Or pretend not to.
If you want to look intelligent, at least show that you understand the other side's point. Then explain, if you don't agree with it, where the argument fails. Do not misrepresent the other side's position.
But what can we expect after all these years.
Already done. Multiple times. And I get exactly what I expect from you. Vacuous arguments. Is there a class in that in law school? I have asked quite clearly and distinctly, "cite ANY evidence that suggests that a knight is the max upper bound and a handicap that large is enough to guarantee a GM win no matter how strong programs get in the future." There is simply ZERO data to support that. There is a mathematical curve that suggests it is false. 50 years ago a GM could give a queen advantage. Today a computer can give a pawn advantage. The slope of that line is quite distinct, and positive.
Feel free to give some actual evidence to support such a claim. AGAIN, not even Larry claims that, except that for programs/hardware today a knight is too much. Barely.
I do not agree that there is a mathematical curve that suggests that programs will be able to win against GM's without a knight.
All the matches with pawn handicap were faster time control than 120/40 so we even did not see that computers today can win against GM's with pawn handicap.
Maybe Komodo can do it but I have no proof for it.
That is my point. We have no proof about anything here, other than today's programs have a chance with a pawn handicap. Saying a knight is the upper bound is completely based on zero information.
bnemias wrote:I should know better than to jump into this... but I'm curious.
Doesn't an ELO difference of X indicate a particular chance of loss or draw regardless of the actual numbers? And if so, where is the breakdown of just comparing ELO for this handicap?
a) perhaps my understanding is off
b) perhaps ELO doesn't apply to handicap games, just regular chess
c) perhaps we lack ELO alignment between computers and humans
A couple are making the assumption that suddenly Elo is meaningless when talking about odds games, that somehow a player with a rating of 2700 in normal chess has an infinite rating when a knight ahead. I don't see any justification for such an assumption, however.
bob wrote:WHAT empirical data suggests that there is an upper bound on this? What specific theoretical proof suggests that there is an upper bound on this?
The mathematical proof is this:
With maximum odds (1 Queen, 2 Rooks, 2 Bishops, 2 Knights and 8 Pawns) the handicapped side just has a bare King against a full FIDE army. No way the handicapped side is going to score any points even when he defends perfectly even against a patzer, or in fact a novice, when the latter has had 10 min of instruction explaining him what stalemate is, and how you can checkmate with two Rooks. (He wouldn't even have to bother learning how Queens, Bishops or Knights move.)
So at maximum odds even perfect play is of no use against a patzer.
No computer, present or future, will be able to play better then perfect.
GMs will not perform worse than patzers.
=> At maximum odds, no computer will ever be able to beat a GM.
At zero odds GMs already lose badly to computers.
So the set of odds where any future computer will not be able to overcome the odds against a GM is non-empty, and the complement of that set also is non-empty.
In addition, the possible odds form a finite set.
=> the set of odds that can be overcome by a stronger player has an element in it where the odds is larger than any other odds in it, but still smaller than maximum odds.
Q.E.D.
Obviously k vs original pieces is the "upper bound". But what is the lower bound, the smallest odds one can give a computer and still be guaranteed that the computer will win? THAT is the question, and my believe is "we have no idea." Some have arbitrarily established this as knight odds, claiming that a GM can play perfectly with knight odds, even though he is nowhere near perfect in even material games.
bnemias wrote:I should know better than to jump into this... but I'm curious.
Doesn't an ELO difference of X indicate a particular chance of loss or draw regardless of the actual numbers? And if so, where is the breakdown of just comparing ELO for this handicap?
a) perhaps my understanding is off
b) perhaps ELO doesn't apply to handicap games, just regular chess
c) perhaps we lack ELO alignment between computers and humans
A couple are making the assumption that suddenly Elo is meaningless when talking about odds games, that somehow a player with a rating of 2700 in normal chess has an infinite rating when a knight ahead. I don't see any justification for such an assumption, however.
The question is how much improvement can we get with knight odd thanks to better hardware.
What is the difference in rating between Komodo 24 cores without a knight and Komodo 1 core without a knight(when both play against chess engines).
I do not have komodo and do not have 24 core machine but I guess that we are going to see less than 50 elo difference against engines if we use 10 minutes/40 moves or slower time control when the difference between 1 core and 24 cores is more than 200 elo in normal chess.
bob wrote:Obviously k vs original pieces is the "upper bound". But what is the lower bound, the smallest odds one can give a computer and still be guaranteed that the computer will win?
Even all the original pieces against just a king isn't a guaranteed win. If the player get sick and can't continue it will be ruled a draw.
Uri Blass wrote:The question is how much improvement can we get with knight odd thanks to better hardware.
What is the difference in rating between Komodo 24 cores without a knight and Komodo 1 core without a knight(when both play against chess engines).
I do not have komodo and do not have 24 core machine but I guess that we are going to see less than 50 elo difference against engines if we use 10 minutes/40 moves or slower time control when the difference between 1 core and 24 cores is more than 200 elo in normal chess.
When playing against itself with knight or against Stockfish with knight, it is not clear to me that "Komodo without knight" will score any points, whether it uses 1 core or 24 cores (in particular at 10 minutes/40 moves or longer).
Assuming there are occasional draws, 24-core Komodo might on the average last a bit longer than 1-core Komodo and therefore give its opponent a few more opportunities to blunder. And if the opponent blunders, 24-core Komodo will be more likely to convert. So 24-core Komodo would then score a bit better than 1-core Komodo. But I fully agree that the "knight odds" Elo difference between the two will be much smaller than the regular Elo difference, as the difference will only be made on a small subset of games.
Uri Blass wrote:The question is how much improvement can we get with knight odd thanks to better hardware.
What is the difference in rating between Komodo 24 cores without a knight and Komodo 1 core without a knight(when both play against chess engines).
I do not have komodo and do not have 24 core machine but I guess that we are going to see less than 50 elo difference against engines if we use 10 minutes/40 moves or slower time control when the difference between 1 core and 24 cores is more than 200 elo in normal chess.
When playing against itself with knight or against Stockfish with knight, it is not clear to me that "Komodo without knight" will score any points, whether it uses 1 core or 24 cores (in particular at 10 minutes/40 moves or longer).
Assuming there are occasional draws, 24-core Komodo might on the average last a bit longer than 1-core Komodo and therefore give its opponent a few more opportunities to blunder. And if the opponent blunders, 24-core Komodo will be more likely to convert. So 24-core Komodo would then score a bit better than 1-core Komodo. But I fully agree that the "knight odds" Elo difference between the two will be much smaller than the regular Elo difference, as the difference will only be made on a small subset of games.
My idea was to give 24 core Komodo and 1 core Komodo to play against weak engines and not against komodo so they may win or draw part of the games.
bob wrote:
I don't know that the machine will be able to win, but there is absolutely ZERO evidence to support that it will always lose. Absolutely ZERO. To make such a claim, there must be something to support it.
what is the reason that this statement of larry would not be evidence that the machine will lose to knight handicap ?
is it because he has not proven that K/C ratio cannot be 64%, or something else. ?
Let's say that there is some handicap that would produce an even score in a serious match between Komodo and Magnus Carlsen. I would estimate that this handicap would be in the 1 to 1.5 pawn range based on Komodo's eval after a long think. Let's say 1.25. Now both Komodo and Carlsen make errors of some average magnitude. We'll call Carlsen's error rate C, and Komodo's K. I think it's pretty obvious that K is much less than C, let's say K = .4xC. If a future engine drops the error rate to zero, then C - K increaases to 5/3 of it's former value, so the proper handicap should also increase in that ratio. That would put it at 208, a bit over two pawns but way below the roughly 3.5 value of knight odds. Of course there is a lot of uncertainty in the above, but I don't think the estimate would be way off. In order for the estimate to reach knight odds, the K/C ratio would have to be about 64%, which does not seem plausible to me
bnemias wrote:I should know better than to jump into this... but I'm curious.
Doesn't an ELO difference of X indicate a particular chance of loss or draw regardless of the actual numbers? And if so, where is the breakdown of just comparing ELO for this handicap?
a) perhaps my understanding is off
b) perhaps ELO doesn't apply to handicap games, just regular chess
c) perhaps we lack ELO alignment between computers and humans
A couple are making the assumption that suddenly Elo is meaningless when talking about odds games, that somehow a player with a rating of 2700 in normal chess has an infinite rating when a knight ahead. I don't see any justification for such an assumption, however.
The question is how much improvement can we get with knight odd thanks to better hardware.
What is the difference in rating between Komodo 24 cores without a knight and Komodo 1 core without a knight(when both play against chess engines).
I do not have komodo and do not have 24 core machine but I guess that we are going to see less than 50 elo difference against engines if we use 10 minutes/40 moves or slower time control when the difference between 1 core and 24 cores is more than 200 elo in normal chess.
You are thinking way too small. Crafty, today, hits 100M nodes per second on my 20 core box. We are talking what about something that is 10,000,000 times faster, or even more. NOT what is possible today. The original question was what about 32 piece endgame tables. Could a computer beat a gm with knight odds or better given either 32 piece eg tables, or such incredible search depth that it plays such positions perfectly anyway. The hypothesis has been offered that once a GM is a knight ahead, he can play _perfectly_ and no machine will ever be able to beat him. I don't see any evidence to support that. For my 7 orders of magnitude improvement in speed, that is close to 24 doublings. That's not going to be 50 or 100 Elo improvement. And that is just the beginning.
I think both computers and chess programs have a long way to go before things flatten out. I'm not sure a knight will be impossible. I don't know, but I certainly know there is no evidence to suggest any sort of limit like that exists. Yes a queen and two rooks is enough and then some. But a knight. Knight and pawn? Rook? impossible to guess.