Author: Albert Silver
Date: 07:17:16 10/23/00
Go up one level in this thread
On October 23, 2000 at 00:24:54, Christophe Theron wrote:
>On October 22, 2000 at 15:23:30, Ratko V Tomic wrote:
>
>>> I cannot think of any other purpose for all this other than to improve
>>> the program's rating in the SSDF.
>>
>>There is no harm all by itself in being at the top of SSDF.
>>
>>> You can't possibly use this for
>>> human players, and you aren't talking about offering playing styles,
>>> as that is already rampant.
>>
>>Human players routinely adjust their play to exploit better the stylistic or
>>capabilty weaknesses of their opponents, programs or humans. Programs aren't as
>>good (yet) at this kind of 'outside the board' strategizing, but they will get
>>better.
>
>
>
>I believe that the current chess programs/hardware combinations are already at
>the level or even better than GMs "on the board". Don't shoot at me right now.
It's true that the performances that have been registered certainly show that,
so if that is true for all to see, why are there still many who insist it isn't
so? Hard-headedness? I think the issue isn't one of Elo, but of how one
perceives GMs. At first, with the gap being much much larger, people simply
pointed to the ratings, afterall, why seek further discussions if this basic one
was sufficient? The issue is in what a GM knows and applies at the board. Again,
traditionally, the game ends up being divided into a number of sections:
openings, middlegame (tactics and position), and the endgame. Computers have
fought the issue of openings by coming with opening books that often range in
the millions of moves. As positional play had been basically reduced to the same
tips handed over to beginners (put a rook on the open file, castle quickly,
develop your pieces ASAP, etc...), tactics became the great equalizer. True,
players were often capable of calculating specific lines much deeper (still
true) and create elaborate positional plans, but the computer fought back with
our very own mistakes. In fact this is a bit why many players hate playing
against programs: they don't feel their opponent played a better game, they just
lost because they made a stupid mistake which was revealed with the greatest of
ease (perhaps also making them feel slightly stupid). Computers have shown one
thing extremely well, and it has been a very valuable lesson: do not
underestimate for an instant the importance of tactics. From this lesson, some
have learned, and some have not. Kasparov himself has acknowledged the change in
perception and some GMs like Yermolinsky (in his masterpiece released last year)
have come to terms and said that there is no correct or 'better' style of play.
So that according to this, programs are indeed playing on equal terms (same
strength), but with different weapons. Then you have a different perception
shown by players such as Seirawan, who published in his book on the 1992
Fischer-Spassky match, that players fell into 4 categories (he had written this
to talk about Fischer's game but it is very revealing):
- Tactical players
- Positional players
- Strategic players
- Builders
The last 3 (!) represent different levels of positional profficiency, and he
succintly declares that Tactical players are by far the weakest of the lot. The
reason is that prior to programs, there has never been such a lopsided player:
3000 Elo in open tactical positions, 2500-2600 in mixed positions, and 2100-2200
(if that) in closed positional games. That isn't an exaggeration in most cases.
Again, the problem is that GMs don't make certain mistakes (they may blunder,
but certain conceptual mistakes will almost NEVER happen) and when the newly
touted 'GM's do this, they are pointed out to be not GMs. I understand.
I myself hesitate greatly to describe them as being anything other than GM
'strength', meaning they aren't GMs, but they're just as strong overall. Mind
you, this same reason is why GMs love computers as well: it isn't a peer, it's
something else, and as such presents a brilliantly challenging opponent to work
on specific aspects of their game in a way a colleague never could.
>
>But the GMs are much much better at taking advantage of the "off the board"
>knowledge (they know their opponent and can adapt their strategy to it).
>
>This "off the board" knowledge and ability to adapt gives them an serious edge
>over computers. There is a big gap to close for computers because of this. It is
>maybe 50 elo points at this time (maybe even 100 elo point for a few
>individuals, but there are not many of them), and it might well be like the
>horizon. No matter how close you get, the horizon moves away.
Yes, but you don't say 'how' it is they are adapting their knowledge. They are
merely emphasizing their strengths (strategy) as opposed to their weaknesses
(tactics) relative to the computer. One solution is to have the program do the
same, hence the 'anti-human' options that are appearing, and the other longer
route, is to try to balance it out and try to find a way to teach it to play
positionally. So far, this has been the single-minded direction followed, but
now we see there are other aspects such as the attack.
BTW, Gambit Tiger has been making so many waves, because it has been showing
another poorly exploited (by computers) aspect of the game: the attack. Since
the attack has always been a domain of tactical players, once programs were so
efficient tactically, it seems that it was overlooked that tactics are only a
part of an attack, not the attack itself. Add to that, the fact that it doesn't
wait for the opponent to make a mistake, but rather tries to win the game by its
own merits, and you have a very entertaining opponent. :-)
>
>Increasing the computers speed will not solve the problem immediately. The gap
>should be closed by introducing some opponent knowledge into the programs
>(opponent modeling).
This may possibly work, but I think that introducing the 'opponent's knowledge'
as opposed to 'opponent knowledge' is more interesting.
Albert
>
>
>
> Christophe
>
>
>
>
>>> Creating a number of personalities to defeat specific opponents would
>>> require a lot of work and would only be of interest to the
>>> programmer. Why would I want to have a Rebel especially geared to
>>> defeat GM Svidler for example?
>>
>>I wasn't talking about specific opponent but only a specific relative advantage,
>>no matter who the opponent is (i.e. the program is being tuned to a set of
>>opponents with particular strength range in particular types of positions). The
>>usefulness of this is surely greater than just better SSDF rating. After all if
>>the program adjusts itself to maximize its strength when playing against you (or
>>whoever else you pit it against), for all you care you're getting a stronger
>>sparring partner, as if you bought a faster computer and used the same program
>>without self-tuning.
>>
>>GT is quite effective against programs which don't have as good king-attack
>>algorithms, more so than other good king attackers, including regular Tiger,
>>probably because it does take risks. The proper degree of risk taking ("proper"
>>meaning as empirically determined) is the performance maximizing strategy. By
>>taking risks, GT makes positions in which it excells more frequent. While
>>regular Tiger may be as good, even better, once the king attack is under way, by
>>virtue of taking fewer risks it won't find itself in a king-attack as often as
>>GT will. To maximize performance (rating) it is not enough to have an edge in
>>some area. One has also to try maximizing the frequency of positions in which
>>one can use that edge, even if some risks are involved.
>>
>>Now, once you agree that it is reasonable to take some risks in order to
>>increase frequency of positions in which one will have an edge, and since the
>>degree of advantage varies for different opponents, the immediate consequence is
>>that one could do even better if the amount of risk-taking is tuned to the
>>specific degree of the advantage one has against the current opponent. And that
>>can be learned in a longer match or hand input if the operator knows the
>>opponent. For example, GT might perform better against regular Tiger if it were
>>to play with risk-taking "instinct" minimized.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.