Author: Steven Edwards
Date: 01:17:06 11/05/03
Go up one level in this thread
On November 05, 2003 at 00:03:06, Peter Kappler wrote: >On November 04, 2003 at 19:24:25, Steven Edwards wrote: >>On November 04, 2003 at 18:47:02, Peter Kappler wrote: >>>Are you aware of the CCT tournaments that have been played on ICC? We've had 5 >>>of these since 2000, with increasingly good turnouts each year. The last event >>>had 45 entries. >> >>Yes. But not everyone has an ICC membership. They are not free. I want this >>to be a free event without any commercial interests upon the part of the >>organizer. The entrants will be playing solely for their own benefit, not for >>spectators who've paid an organizer who in turn does not pay the entrants. I note that there is no suitable reply to this concern. >>>1) Why limit participation to 24 programs? I think it's much better to take as >>>many entries as possible, even if it means playing a Swiss instead of a >>>round-robin. >> >>A round robin event produces a much clearer picture of the relative playing >>strengths of the entrants. The limit of 24 entrants should be sufficient to >>cover all the active authors with a few slots in reserve. > >Really? How did you determine this? Be careful when making assumptions. I suppose in part it depends on the definition of "activity". Note that I said active authors, not active programs. >> All of the great human tounaments have been round robins. > >That won't be much consolation to author #25. And if the limit were 1,000, it would be no consolation to author #1,001. A line has to be drawn somewhere, just as a line is drawn for human tournaments of fixed size. >>>2) I hope I'm wrong, but it sounds like you are planning to develop a server for >>>this. Why? ICC works perfectly fine, and they are willing to give free >>>accounts to program authors. My program already understands 3 protocols: >>>Winboard, UCI, and ICC. I'm not excited about adding support for another. >> >>I wrote "well documented". I didn't say "new". Be careful when making >>assumptions. > >Fine by me, as long as whatever you use support existing protocols. I won't add >support for a new protocol just to play in one event. It may be that FICS has to be extended slightly to support active pairing where the server initiates contact with the programs instead of vice versa. This is still up in the air at this (early) point in time. Ten years ago, some authors refused to add chess data standards support to their programs saying pretty much the same as you've just written. This is not an argument for change for the sake of change itself, but it's a good argument against an unconditional policy of no changes at all. >>Not all program authors have a free ICC account. In the ICC documentation, a >>"free" program account is given only if the author already has paid for a >>regular account. > >No, they've waived this rule for the CCT events. Since that's the case, I hope >you will consider using ICC. It certainly hasn't be waived for authors who only want to test their program on ICC. Also, see my earlier objection to the commercial nature of ICC with respect to what is to be a non commercial scientific project. >>The server used will be a dual P3 Intel rackmount running RH9 Linux. A backup >>AMD single CPU will be available. Unlike the ICC, the server will be run >>exclusively for the entrants so there should be plenty of CPU power available. >>The server IP and passwords will be restricted to the entrants so there should >>be little chance of excess loading. > >While this is nice, it's not like ICC has any issues with CPU load. Some ICC users will disagree here, although in some of the cases it may be bandwidth limitations associated with thousands of active connections with a significant portion of the load used for chat. >>>3) One game per day in the middle of the day? Most of us have day jobs, so we >>>won't be able to watch the games. One of the great things about the CCT >>>tourneys is the weekend format makes it possible for the programmers to be >>>online, chatting and commenting during the games. This is loads of fun - and >>>the main reason for participating in these tourneys, IMO. >> >>One of the tests of the entrants is the ability to play autonomously. There is >>no one time period during the day that will please everyone. > >Yes, autonomous play is a requirement at the CCT tourneys, too. And you're >right, there is no ideal time of day, but clearly the best days of the week are >Saturday and Sunday when more people can watch and discuss the games, live. Saturday and Sunday are playing days in the announced schedule. I am planning some sort of facility where those at work can peek into the server (if their resources permit) to get a real tme view of things. Also, a chat facility can be set up, although it does not necessarily have to be set up on the tournament server. >>I will make myself available for technical assistance during all scheduled >>playing time plus for at least two hours prior to the start of play each day. I >>assure all entrants that running a smooth event will be my only priority; I have >>no commercial chess interests at all. > >Nor do any of us who have organized the CCT events. ICC is a commercial interest. It has responsibilities towards many, many users. I can dedicate all of my resources to only a couple of dozen authors. Furthermore, I hope to be able to continue this free service to authors: a chess server dedicated for their use and free from loading by non authors. >>>4) Why the title "North American" championship? What's North American about it >>>other than possibly the location of the server and the home address of the >>>organizer? >> >>It has been a long time (nine years?) since there has been an exclusively North >>American event. It is intended, in part, as a prelude to physical North >>American events. These will be an alternative to the European events that, >>while certainly worthwhile, do little to stimulate CC activity in North America. > >So entry is restricted to programs authored in North America? If so, that's a >shame. It will seriously weaken the event. Some may take your last comment as an insult while others will take it as a challenge.
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.