Theoretically, with over 3100 elo, it could sweep easily all the humans and win without too much problems - something of laugh -, but in fact, I am not so sure this can happen. I have some doubts the program can finish a tournament like this with that incredible elo performance, or at least that it can always maintain this overall elo-performance in other several and quite strong human tournaments.
Do you really think that Rybka can play always at that high level against the strongest humans in the world, and also to maintain constantly their elo-performance in other tournaments?
If I was a very strong human at the Linares tournament, I would refuse to play Rybka, since I came to play against other humans, so having to play a computer would just disrupt my play (or strategy, or would distract me, tire me etc.)
Now, Imagine that humans cannot refuse to play against Rybka and they always agree to give their best try in those tournaments where Rybka will participate.
Could Rybka maintain their 300 points of difference over Anand, after several quite strong human tournaments?
> Could Rybka maintain their 300 points of difference over Anand, after several quite strong human tournaments?
In that hypothetical situation I don't think so, but (and I'll repeat myself) it's a problem with the ELO formula itself, as it doesn't manage 300 ratings points of difference properly (So, the stronger player has a great trouble trying to perform at its expected ELO, but against opposition of about the same level, the task is easier.)
If the sponsors decide that the prizes are the same and the players have to play against rybka then
I see no reason for human players to refuse to play.
If 2700 players agree to play X games against humans for expected money of f(X)$ then I see no reason for them to refuse to play X-1 games against humans and 1 game against rybka for expected money of f(X)$
Maybe it was a good reason in the Junior case but it is not a good reason today.
The computer is a clear favourite to win and humans can do best by spending more energy about the games against humans
and not making special preperation against the machine.
"their game against Junior affected their play against the human participants."
Not convincing accuse.
I assume humans do their best and if preparing against the computer is going to cause them problem against humans then they will not prepare against the computer.
"If Rybka played in Linares, it would most likely win the tournament, which would probably diminish the 'human winner's' achievement in public perception."
The last sentence is the only logical reason that I see for GM's to refuse to play against rybka but this is based on the assumption that people are stupid and expect rybka not to win the tournament by a big margin.
I see no reason that is not about money to refuse to play against rybka and this reason is clearly about money because maybe GM's are afraid that if rybka wins by a big margin the prizes for humans are going to be smaller in the future.
Well, if this is what they'll do, then what's the point of such a tournament? Additionally, this reasoning only shows why such a tournament will be really expensive. It is quite embarassing to be butchered by a computer, and this is just what can be expected if you are not prepared (although it's hard to avoid with serious preparation, too). Kramnik can make a lot of money by playing computers AND demand match conditions which clearly favor him. How much money do you think will convince him to play a game where he knows he does not stand a chance? Many world class players have big egos ...
"[...] that people are stupid and expect rybka not to win the tournament by a big margin."
Many people are still not aware of the superiority of computers compared to humans. No need to call this stupid. But even if everybody is aware of this: The coverage of the tournament will most likely focus on the winner of the tournament, and report less about the 'best human'. I don't think it's irrational if the best players expect to get the attention they (think they) deserve.
Of course the prize money would have to go up, and not by a small amount. In effect the players would be selling a piece of their reputations if they were getting thrashed by a machine and surrendering their traditional preeminence in these tournaments. What's more, human chess generally would be diminished.
If the knowledge that a $2000 computer equipped with $50 software could resoundingly thrash the top grandmasters time after time permeated the global chess community I have to think it would be subversive to the game generally. Look at how subversive it is when unrated players equipped with some hardware can routinely beat grandmasters aided by computers! All of these trends tend to dilute the value of human skill.
> All of these trends tend to dilute the value of human skill.
Skill at chess, that is. Contrariwise, they show the power of human intellect to concoct a suitable out-sourcing agent, capable of handling most of their chess concerns.
Feeling bad because a human cannot compete with a computer is like feeling bad because an army equipped with swords and javelins cannot compete with an army with automatic rifles. There is little place for sentimentality in the march of technological progress.
> I refer the gentleman to the answer I gave a moment ago
You need to put "honourable" before gentleman. With that, I yield back the remainder of my time.
Ok ok I'm putting my Hansard away now.
Using computers for analysis is not the same as playing them.
If you do extra preperation against computers then you look for positions that they do not understand when humans use computer to analyze line that they play that are usually not positions that programs do not understand.
The GM does not know it from experience.
The GM is using computer to analyze lines that he plays in practical games and not to analyze lines that he never plays and the computer is relatively weak at them.
Suppose that the GM can perform at level of 2700 in case of learning the line 1.e4 c5
Suppose that the GM can perform at level of 2600 in case of learning the line 1.e4 e6
Suppose that the computer can perform at level of 3000 in case of learning the line 1.e4 c5
Suppose that the computer can perfrom at level of 2600 in case of learning the line 1.e4 e6
It is a bad idea for the GM to invest time on 1.e4 e6 in case of playing against humans when
it is a good idea for the GM to learn this line if he prepares against the computer
But of course chess is a bussiness a little one compare to others but a business so a probable score for a progtam like 12/14 would be katastrof.
By the way i think the worst moment for computer chess history was Hydra performance over Adams. Humans don't let that happen again easily. If Hydra could do it once can be again and thats enough of a proof.
Of course a match with Hydra, although highly improbable, could be an indirect indicate.
"As a personal opinion, and a little bit out of topic, I think it would be very interesting to see Rybka playing in a very strong human as Linares chess tournament where the best humans are participating. This can clear some doubts about what an entity can do with that incredible and eloquent colossal strength. "
The elo of Rybka is about engines chess, but with the top human of chess, this eloquent fall in value.
The match with Dzindzi shows that you can win and above all tie, although it with a pawn advantage.
> although it with a pawn advantage.
Don't forget it was Pawn+Move.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill