Author: Ed Trice
Date: 22:33:36 12/31/03
I hope nobody minds that I spawn a new post although this is really part of a threat. I was getting carpal tunnel scrolling down and follwing the nested, nested, nested replies. I will try to get everyone up to speed quickly. First to all those "chess only" readers, this is Gothic Chess, played on an 80-square board with two new pieces; the Archbishop (Bishop + Knight) and Chancellor (Rook + Knight). If you are not interested in this variant, stop reading now :) Ok, if you are still reading and never heard of Gothic Chess, a pitstop to the link http://www.GothicChess.org will be worth the time. A discussion had ensued about my Gothic Vortex program (80-bit using 64-bit and 16-bit components) and the performance of 32-bit chess programs. I think I can distill the essentials of what could be viewed as a disagreement or disbelief among the programming enthusiasts. First, Gothic Vortex borrows some of the Craft technology, thank you Dr. Robert Hyatt, my new best friend :) Second, during the construction of the move generator, I reported a nodes/second count of 140 million, causing some consternation. It should be emphasized that this figure was not from an alpha-beta search from a fixed position, nor anything similar. I had mentioned how I had arrived at this figure, by putting pieces on each square (from a1 to j8) then repeatedly calling my "move_gen" routine in a for loop a fixed number of times (I think 100,000 per position) then counting the nodes spawned, then dividing by the time taken. I understood this was not a representation of the full speed of the program; as it was in the early stages, it was the only metric I had to test my 80-bit move generation routines. Someone had contacted me asking me for test data. I provided what I had. Thirdly, I could only compare my chess variant of Gothic Chess against the only other Gothic Chess program out there, which was from the Zillions-Of-Games company. Granted, this is an engine that can work for games of ANY size and rule configuration, which is a feat unto itself, but my results comparing it to Gothic Vortex were veritable. Gothic Vortex is 30 times as fast as this generic engine. It is the only way we can compare engines, as no other chess engine operates on 80 squares. Gothic Vortex played a 10 game match with Zillions and won 10-0, not even a single draw, at 3 minutes per move. Since Zillions plays the same way every time (no opening book) we let Vortex pick openings at random (it has a book of about 6000 opening positions.) Now several "proposals" have been offered to compare the engine speed of Gothic Vortex to other chess engines, but comparing 80 square boardds to 64 square boards seem rather pointless. So, here is the kicker. I will pay anyone $10,000 that produces a Gothic Chess program that can defeat Gothic Vortex in a match to take place in the city of Philadelphia no later than August 1, 2004. I would like to know if there are any takers. We can then decide the length of the match, and the format. Please note: this is not an idle boast. I am interested in knowing if those who are talented chess programmers would believe it worth their time and effort to try and shoot for this mark. I am open to negotiating the details.
This page took 0.02 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.