Computer Chess Club Archives


Search

Terms

Messages

Subject: Some thoughts on draws...

Author: Russell Reagan

Date: 16:37:27 02/07/03


Regarding draws in human vs. computer events...

Kasparov said something that made sense, and got me thinking. He said that
Junior was never able to "out play" him (which I took to mean strategically),
and that it relied on Kasparov's blunder to win its game. Kasparov won his game
by playing a solid game from start to finish.

So let's establish a few pseudo-facts. Computers have a difficult time "out
playing" the top humans strategically. Computers WILL hammer you if you make a
mistake. The top humans are tactically sound for the most part, but still make
some mistakes.

These pesudo-facts lead me to the following conclusion. Since computers have a
difficult time "forcing" a win from a strong player, when no blunder is made,
this leads to either draws or losses for the computers, with the occassional
win. Humans know that one mistake will almost always cost the game against the
computer, leading to an attitude of being content with a draw, even when winning
chances exist. Final outcome, lots of draws, with wins scattered here and there.
So far, this makes more sense to me than all of the conspiracy theories that
people like to throw around.

Regarding draws in general...

I would like to see some method to make draws less common, or at least harder to
arrive at. I believe most people would agree, if it could be done in a good way,
that didn't take away from chess or change the game too much. So what is the
main problem? The main problem, as I see it, is not draws by repetition, or by
the 50 move rule, but by agreement. Very often in games between strong
participants (human or computer), you see draws agreed upon in under 30 moves.
Let's take a look at the different kinds of draws.

I see no problem with draws by repetition. They don't waste a great deal of
time, and it is fairly clear that the game would go on in the same manner
forever, so a draw is good.

The fifty move rule has a good intention, but it's not my favorite. It takes far
too long to actually arrive at the 50 move rule, and it also alters the
theoretical game result of some endings that are known forced mates, but which
are declared draws by the 50 move rule because it incorrectly assumes that no
progress will ever be made. Perhaps adding checks to the list of "captures or
pawn moves" would make the 50 move rule better?

I don't see anything wrong with stalemate, and this is not really the source of
many draws anyway in practical play.

Finally, draw by agreement. Here is where I see a problem. I think there could
be some type of altering of the drawing rules that wouldn't alter the game of
chess a great deal. We have a maximum moves without progress rule, the 50 move
rule. This has flaws, but is not so bad, and without it you might have bigger
problems of stubborn mules playing on for hours upon hours in a dead drawn
position. How about a minimum moves without progress before a draw can be agreed
upon? I think it would be nice to have a rule that required there to be no pawn
moves or captures for (say) 10 moves by each side (20 ply) before a draw can be
agreed upon. The exact number may be higher or lower, but I think this would do
a lot of good for the "drawing problem" that makes potentially exciting games
into boring snoozers. There could be additional qualifications, such as, if
there are good captures (captures of a piece of less or equal value compared to
the capturing piece) then no draw by agreement, or other things that can cause a
great shift in the game. Perhaps the "good captures" addition is taking it too
far, but I think the point is clear. There should be some lower bounds on move
number that must be reached, or number of moves without progress before a game
can be drawn by agreement.

So what do people think about this? Is this just another flawed attempt to spice
up the game? Or does it have some potential? I doesn't really change any rules
that affect the theoretical outcome of the game (like the 50 move rule does), so
I think it is at least better than some suggestions you hear about this problem.



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.