Author: Vincent Lejeune
Date: 14:22:14 04/10/02
Go up one level in this thread
On April 10, 2002 at 16:22:18, Robert Hyatt wrote:
>On April 09, 2002 at 16:02:47, Roy Eassa wrote:
>
>>
>>Let's see what statements BOTH sides can agree on:
>>
>>1) In most highly open, tactical positions, the strongest computers are usually
>>stronger than even the strongest GMs.
>>
>>2) In many more-closed positions the strongest GMs are stronger than any
>>computers.
>>
>>3) A GM can maximize his chances and thus minimize the computer's chances by
>>avoiding the types of positions in #1 and creating those in #2. THIS IS A SKILL
>>UNTO ITSELF.
>
>
>Here is a cute question:
>
>We are going to play a game where each of us (two player game) has a coin.
>I can show you either a head or a tail, and you do the same to me. We both
>show our coins simultaneously. If we both show heads, you owe me $1. If we
>both show tails, you owe me $3. If we show different (head for me tail for you
>or vice-versa) I pay you $2.
>
>Do you play this game with me?
>
>(Hint: it looks evenly matched but it favors me)
me : tail(25%) --> - 1 for me
/
you : tail(50%) --
/ \
/ me : head(25%) --> + 2 for me
---
\ me : tail(25%) --> + 2 for me
\ /
you : head(50%) --
\
me : head(25%) --> - 3 for me
conclusion : wining expectancy : 0 !!!
Where's the flaw ???
>
>This is the situation with GM players vs Computers. If they know how to
>unbalance the game then they increase their chances. If they "play chess"
>then they will find problems..
>
>
>
>
>
>>
>>4) The skill described in #3 is a somewhat DIFFERENT one from that which each GM
>>has focused on over his lifetime. Not completely different, but certainly not
>>identical either.
>>
>>5) Nowhere near as much time has been spent by humans over the centuries at the
>>skill described in #3. Besides the fact that it is obviously a very NEW skill,
>>historically speaking, there also has not been a financial incentive for
>>spending YEARS OF HARD WORK (like GMs do with traditional chess skills)
>>developing this new skill.
>>
>>6) There are some non-GMs that apparently have the new skill in greater degrees
>>than the top GMs appear to. This is probably true because there is a far larger
>>sample size of non-GMs than GMs in the world and because few GMs can afford
>>(money-wise) to divert their attention to this new skill.
>
>I think the difference is for another reason. GMs got where they were by
>playing their own brand of chess. And they have perfected it to a level that
>most only dream of. Lower-rated players are still looking for that "niche"
>they can play in and do well, and by "thinking outside the box" they manage
>to give computers a lot of trouble... The GMs, however, want to maintain
>their "edge" without diving into new waters, so they tend to play normal chess
>and have their troubles...
>
>
>
>
>>
>>
>>One could reasonably conclude that today's GMs are are simply quite WEAK at this
>>new skill THUS FAR.
>>
>>In order to consistently beat top computers, a human will require BOTH great
>>chess skill and ALSO high "avoid heavy tactics" skill. One without the other
>>will probably not lead to a human consistently beating the top computers.
>>
>>It will be interesting, IMHO, to watch the race: GMs improving this new skill as
>>the years go on (the best ones probably can't improve their traditional chess
>>skill very quickly any more), versus computers getting faster and "smarter."
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.