Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Scrappy #1 at ICC

Author: Robert Hyatt

Date: 15:19:48 07/10/01

Go up one level in this thread


On July 10, 2001 at 17:12:00, Albert Silver wrote:

>On July 10, 2001 at 16:07:42, Robert Hyatt wrote:
>
>>On July 10, 2001 at 11:41:33, Albert Silver wrote:
>>
>>>On July 10, 2001 at 11:09:39, Robert Hyatt wrote:
>>>
>>>>On July 10, 2001 at 08:50:20, Chessfun wrote:
>>>>
>>>>>On July 10, 2001 at 08:48:09, Chessfun wrote:
>>>>>
>>>>>>Since the Spitfire message, why not a Scrappy one.
>>>>>>
>>>>>>Seems Spitfire caught the draw bug.
>>>>>>
>>>>>>Recent games of SpitFire:
>>>>>
>>>>>Opps
>>>>>77: - 3379 W 3223 Mecking      [ br  5   7] B22 Mat Jul 10 01 08:24
>>>>
>>>>
>>>>Mecking can certainly be a problem when he wants to be.  :)
>>>>
>>>>Bob
>>>
>>>I just want to take advantage of this moment to congratulate you on your great
>>>achievement. I realize it has been a long and hard climb as you aspired
>>>incessantly to the highest ranks. And now, after all your efforts, and having
>>>paid your dues (i.e. the last installment on the Quad), YOU (= Quad + Crafty +
>>>breathtaking formula) have finally reached that most coveted of places in the
>>>sun: the ICC blitz no. 1 !! Yes! Now Crafty can say it is the best of the
>>>best... AGAIN! Brothers, sisters, let us join hands and rejoice!
>>>
>>>Why the very honor of having my modest post linked to yours is one I will
>>>cherish for the rest of my days, oh great
>>>quad-plus-crafty-plus-breathtaking-formula account owner you! Could I humbly ask
>>>for a scanned image of the creditcard receipt that made all this possible? It
>>>would be given a place of honor in the forthcoming site in the "How it all
>>>started" section .
>>>
>>>                                        Albert
>>
>>
>>:)
>>
>>Seems that "scrappy" is a continual cause for conversation.  As I said before,
>>I ran it as an experiment, then I terminated it, and was greeted by lots of
>>messages (and a few emails) asking for "scrappy" to return.  Too many computers
>>continually playing Crafty.  Or if it is free, computer operators jump in as
>>soon as a human game ends.  I gave up on trying to stop this since some are just
>>going to be discourteous no matter what.
>>
>>Scrappy solves the problem.  It plays the kind of opponent I want to see games
>>against (human GMs).  It doesn't play with computers.  And it does, on occasion,
>>sit idle for long periods of time.
>>
>>All ok by me of course...
>>
>>I don't care whether it is ranked #1 or #1000.  Any more than I care about
>>the rating of "crafty".  I only care to see it slowly improve, which is the
>>goal for playing on ICC.  If I decided to stop working on the program for a
>>while, it would disappear from ICC for that time...
>
>I remember when you explained the purpose of the experiment, which by your own
>words is now over. As the purpose is now to simply allow a human-only Crafty
>running, independent of ranking, could you change the formula to allow strong
>humans to play that aren't necessarily in the 400 point range? Such as a cutoff
>range of 2600-2700 up? This would allow less dexterous GMs to have a shot at it
>also. And if they score points, you'll still learn from how they did it.
>
>                                       Albert


I will do this.  I think I will try a rating cap of > 2600 for simplicity.  But
I suspect this will lead to more complaints as a few IMs will play it over and
over trying the usual anti-computer stuff...  With them at 2600 and scrappy at
3400, they have nothing to lose and everything to gain...

And since the 3200-level GMs have their hands full, will they play it if it is
rated 3000 or so?  Probably not, ie they hardly ever play Crafty now...

It's a catch-22.  When winboard protocol version 3 comes out, this will be
fixable as I will be able to limit the number of consecutive games and cut some
of that out.



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.