Computer Chess Club Archives


Search

Terms

Messages

Subject: Hedlund positional testsuite anomaly?

Author: Paul

Date: 11:04:42 03/16/00


Hi all,

Just a few moments ago I concluded running the testsuite on my own
program, albeit with only 1 minute per position. This is because it's
a bit frustrating for me to not be able to use the computer for anything
else for 16*10=160 minutes (yes, I like to program, I'm an addict :).
It's a P2-400 by the way.

Now I already saw Insomniac score pretty well using less than 10 seconds
per position and I also saw the result of Ernst Heinz's program (113),
but my result must be a statistical rarity.

The score is: 129 (5 10 10 7 7 10 10 10 10 10 10 8 0 5 10 7) !!!

No I don't think my program is brilliant, I'm only just starting to
build in some positional knowledge. Probably what's happening is that
the scores of the various moves must be very close in my eval()
function, and by chance Pretz chooses the right ones.
I also don't score nearly as well on other test suites.

But still, it's uncanny, think I'm gonna buy some lottery tickets
tomorrow, Pretz says he can conjure up some great numbers ;)

Have any of you more-experienced-amateurs tried your programs on
the suite? And what are your thoughts on your results and on the
suite? I would be very interested in any replies.

Thanks,
Paul



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.