Author: Dan Ellwein
Date: 07:44:09 03/21/00
Go up one level in this thread
On March 20, 2000 at 23:44:31, Robert Hyatt wrote: >On March 20, 2000 at 06:08:22, James T. Walker wrote: > >>On March 19, 2000 at 04:28:53, Steffen Jakob wrote: >> >>>Hi! >>> >>>Does anybody have an as complete as possible list of handles of original engines >>>which are playing at ICC? I think about allowing - at least temporary - only >>>orginal programs (and some selected clones like "Beadle") to play Hossa, because >>>I saw too many games against clones in the last weeks. >>> >>>Best wishes, >>>Steffen. >> >>Hello Steffen, >>I'm having trouble understanding why programmers on ICC want to limit who their >>programs play. Crafty is complaining about playing too many Chess Tiger clones >>and Scrappy does not want to play computers except at long time controls(If I >>understand the formulas --It's possible I do not). I know Bob wants to maximize >>his programs vs GM's but let's be honest, most GM's are easier to beat than >>programs like Fritz or Tiger or Junior. Is it the rating points that drive >>everything on ICC? Are the rating points so precious that programmers have to >>select who they play in order to maximize the points? > > >No... you miss the point _totally_. > >First, I am completely baffled why everyone wants to run a chess engine on >ICC? I can't imagine a bigger waste of time and resources, for those that >are running engines "just because". Some of us are actively developing the >programs, and use ICC as a test bed. Those of you that are doing this "just >because" are doing two things: (1) you are draining away the very GM players >that I want to play; (2) you fill my game logs with nothing but games vs a >specific program. I implemented the Tiger restriction when someone pointed >out that the last 20 games Crafty had played at one point were against 5 >different Tiger clones. Playing one program over and over is not something I >want to do, as my goal is not to develop an engine tuned to play against Tiger, >but rather to produce an engine tuned to play against IM/GM players. > >I put !computer in scrappy's formula for that reason... it seems that "some" >thought it perfectly ok to play the max games vs crafty, then switch over to >scrappy and do it again. Since I run both, and look at the games, I decided >that I would simply prevent computers from playing scrappy period, and restrict >the number of games vs any single program to no more than 4. > >I really don't care whether anyone likes my policies on ICC or not. I'm not >doing this for anyone but _me_. I find it strange that several people want to >complain about my computer policy, yet they are not working on any engine at >all, but rather are just playing something on ICC that they bought off the >shelf. > >Again, I don't see the point. IE why not buy a copy of Oracle, and just do >the same query over and over and over? Burns just as many computer cycles, >accomplishes just as much for the improvement of Oracle as it does for the >improvement of an engine, and it would ease the program 'glut' on ICC that >has drastically cut down on the number of games legitimate program developers >are seeing each day... > > >> Is everything done on ICC >>to make your program "Look" stronger than it really is? Why do the computer >>operators on ICC seem to take everything personal as if they have been violated >>when their programs lose? Don't you learn more from a loss than from beating >>someone that blundered the game away? > >Rather than asking what _I_ learn, how about explaining what _you_ are >learning? Did you write the code? Are you supporting the author's testing >by sending him all the games/logs so that he can learn what is going on? > > > >> Don't the losses show you more places to >>improve your program than the blunders by opponents? Of course I'm not a >>programmer so I'm having trouble relating to the formulas and finger notes >>designed to limit who plays and how many games can be played. >>Jim Walker > >I'm having a completely different problem understanding all this, myself. What >in the world is so intellectually challenging about running a program written >by someone else? What is the purpose? And why is it that a a legitimate engine >author _must_ either play such clones or else we are somehow behaving badly by >not wanting to play such a clone over and over and over? > >Explain to me where I am thinking wrong here... don't think you are 'thinking wrong here...' you seem to be level-headed about what you are doin'... PilgrimDan
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.