Author: Robert Hyatt
Date: 20:38:27 04/25/00
Go up one level in this thread
On April 25, 2000 at 23:27:40, Dann Corbit wrote: >On April 25, 2000 at 23:06:56, Peter Kappler wrote: > >>On April 25, 2000 at 22:28:04, Robert Hyatt wrote: >> >>>On April 25, 2000 at 21:40:32, Jerry wrote: >>> >>>>Sorry James, check his FI notes and see for yourself. I do notice his last 20 >>>>games are against Goldmund(GM) elo 2553. 16 wins but with a quad xeon and elo of >>>>3300+ that is nothing special. >>>>He SHOULD BE able to play extremely good against both humans and computers! >>> >>> >>>Finger notes mean nothing here. >> >> >>Bob, why not just update Scrappy's finger notes so this issue won't keep coming >>up? Many ICC users don't know about the 'vars' command. > >I don't think you will ever make everyone happy. >Even if the notes were accurate, the bob-haters would demand to know why only >humans are on the list. > >Should a person's life revolve around trying to placate a mob that cannot be >satisfied anyway? > >Dr. Hyatt is gruff and forceful, and collects both admirers and detractors and >that won't change much, if I am guessing right. > >These tedious tirades are one of the things most annoying about moderation. >That's because they are right on the edge. The comments made -- are they over >the line? Are they just self defense reactions? > >You poke Bob, he pokes back. You step on his toe, he steps on yours. People >will use that as a tool to pull him into pointless discussions (though there are >also fruitful discussions where this takes place). I have noticed (over the >years) that as soon as you become fully civil, Bob will become fully civil. >It's actually a pretty interesting trait. One moment he is berating some net >pest for whatever and the next he is patiently explaining how to set up the >crafty.rc file to the same insipid moron. > >At any rate, my two cents is that Dr. Hyatt is going to be himself, and he is >not going to program the way that others want him to or change his procedures to >those that others desire unless he sees a clear benefit for himself. Seems to >me that such is within the sensible guidelines of allowable choice. Yet >sometimes I wonder if he did (for instance) work a few weeks on optimizing >crafty for 'ponderless' operation if it would not actually save time from having >to constantly wrangle about it with those who crave it. I was hoping you were borderline going to volunteer to do that, sort of like Mike did a couple of years ago to tweak with the timing code. He spent a year trying all sorts of ideas, some of which were good, some of which were bad, but we ended up with a fairly simple algorithm that utilizes time reasonably wisely (especially compared to some programs that end up in an endgame with a huge time surplus and a material defecit due to not using enough time to avoid problems.) But he did all the testing and changing on his own, with occasional discussions and test sessions using crafty as well as his ICC account "fitter". The problem is "time". And I _really_ want to make some progress on Beowulf Crafty, although the 'crafty critics' will again howl in protest when it shows up on 10 quad xeons. :) If there was nothing else on my 'to-do' list, I'd probably tackle this to end the discussion once and for all. But my to-do list is very long, and my available time is much shorter. Which is more important: making it play better overall chess; or making it play better in a crippled test mode that is really not going to be used for serious games? I would much rather work to make it do what it did today, which was to win a string of 30+ games without a single draw, vs FIDE players rated 2550 and up. That is _not_ easy. And yes it requires a bit of luck. But it also requires a _lot_ of skill by a program, no matter how lucky it might be. I find that much more interesting to work on rather than a ponder=off match configuration...
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.