S.Taylor wrote:I would simply love to see what happens if a reputable GM had 6 games vs the strongest machine (or the best Rybka), with the machine giving even QUEEN odds to te GM.
Maybe that would seem bad manners as it is like not resigning, in the mere hope that the opponent will make a stupid blunder. And I would feel it unlikely something would not go wrong in 6 games.
However, to make it more closer to reality, how about just rook odds? EVERYONE is no doubt very sure that the GM would win 6-0.
But _I_ would estimate around 4.5-1.5 to the GM.
Atleast, THIS could be the goal to defeat this.
I would consider 5.5 to be a total victory, however we might be surprised how hard he has to fight for the "expected minimum" of 4.5.
(I'm not saying that the GM SHOULD not win 6-0. Maybe with practice in this setting, it would be easier. I'm also not talking about blitz nor long times, necesarily, But atleast 45 min per game).
I believe that with queen odds, it would be 6-0 even against a good IM player, much less a GM. The computers are _not_ that good...
Why a strong I.M.? An expert should be more than adequate.
You probably mean even at rook odds as "adequate", to normally win, and perhaps get 6-0. But with queen odds, more than that (adequate).
But I don't think it should be thought that I myself have not been around 2200-2300 in chess playing level, but this is mainly according to tests. Sometimes went well over 2300, but on average 2200-2300. I even think some tests are exxagerated in what is expected from playes at certain ratings. Raymond Keenes positions from "the times" however, looked a little easy, though he himself has told me over the phone that they were well tested on rated players, before claiming estimates)
S.Taylor wrote:
But I don't think it should be thought that I myself have not been around 2200-2300 in chess playing level, but this is mainly according to tests.
I think that not a single one of all the thousands of chessplayers i met has ever even tried to guess his strength based on "tests". Is it too annoying for you to "play" chess to get some real information about your probably much much lower real strength?
Results in occasional games with rated players also concured. But no, I did not get the oppurtunity to go into rated tournaments at the time when I was at my best (to date). before then, I did a little, but they were not rated in elo.
S.Taylor wrote:I would simply love to see what happens if a reputable GM had 6 games vs the strongest machine (or the best Rybka), with the machine giving even QUEEN odds to te GM.
Maybe that would seem bad manners as it is like not resigning, in the mere hope that the opponent will make a stupid blunder. And I would feel it unlikely something would not go wrong in 6 games.
However, to make it more closer to reality, how about just rook odds? EVERYONE is no doubt very sure that the GM would win 6-0.
But _I_ would estimate around 4.5-1.5 to the GM.
Atleast, THIS could be the goal to defeat this.
I would consider 5.5 to be a total victory, however we might be surprised how hard he has to fight for the "expected minimum" of 4.5.
(I'm not saying that the GM SHOULD not win 6-0. Maybe with practice in this setting, it would be easier. I'm also not talking about blitz nor long times, necesarily, But atleast 45 min per game).
I believe that with queen odds, it would be 6-0 even against a good IM player, much less a GM. The computers are _not_ that good...
Why a strong I.M.? An expert should be more than adequate.
Are you talking about queen odds ? Do you realize what you are saying ? I'm a beginner (~1450 ELO) and I can beat Fruit 2.3.1 really easily at queen odds (I have made some test games yesterday and I win all the games without problem). Giving queen or rook odds to a GM or IM would be an insult for them...
S.Taylor wrote:I would simply love to see what happens if a reputable GM had 6 games vs the strongest machine (or the best Rybka), with the machine giving even QUEEN odds to te GM.
Maybe that would seem bad manners as it is like not resigning, in the mere hope that the opponent will make a stupid blunder. And I would feel it unlikely something would not go wrong in 6 games.
However, to make it more closer to reality, how about just rook odds? EVERYONE is no doubt very sure that the GM would win 6-0.
But _I_ would estimate around 4.5-1.5 to the GM.
Atleast, THIS could be the goal to defeat this.
I would consider 5.5 to be a total victory, however we might be surprised how hard he has to fight for the "expected minimum" of 4.5.
(I'm not saying that the GM SHOULD not win 6-0. Maybe with practice in this setting, it would be easier. I'm also not talking about blitz nor long times, necesarily, But atleast 45 min per game).
I believe that with queen odds, it would be 6-0 even against a good IM player, much less a GM. The computers are _not_ that good...
Why a strong I.M.? An expert should be more than adequate.
Are you talking about queen odds ? Do you realize what you are saying ? I'm a beginner (~1450 ELO) and I can beat Fruit 2.3.1 really easily at queen odds (I have made some test games yesterday and I win all the games without problem). Giving queen or rook odds to a GM or IM would be an insult for them...
I agree that giving a queen or rook odd if we do not talk about time control of 45 seconds per game or something similiar so the human may have problems not to lose on time may be insult to strong players.
Note that fruit is not designed to do best at beating humans with queen odd.
Maybe it is possible to beat 1500 players with queen odd at time control of 45 minutes per game with the right program but I believe no program can beat 2000 players with queen odd at the same time control.
The reason is that even chess programs cannot do impossible things and I believe that even thinking in way of planning the mistakes of the opponent is not going to help against 2000 players.