Author: Dann Corbit
Date: 18:24:11 08/31/01
Go up one level in this thread
On August 31, 2001 at 19:14:12, Jorge Pichard wrote: >After a successful 1st place at the WMCCC for Amir with his excellent program >Deep Junior and for Stefan Meyer with Shredder 5.32, it must be hard to admit >that when pitted against Gambit Tiger II in a series of 40 games such as in the >SSDF, Gambit Tiger II is proving to be stronger than both when using comparable >Hardwares. I think that Amir Ban and Stefan Meyer-Kahlen are very computer saavy. They both know a lot about mathematics and how computer contests work. I would be very much surprised if they feel any turmoil when something unsurprising like a run of SSDF games occurs. On the other hand, I imagine that they feel great consternation about other people's [customer's] unrealistic expectations in connection with these contests. Only one program can top the SSDF. And yet, when you look at the error bars, any of the top three or four programs are really equivalent. Only one program can win the WMCCC [in a given category] and yet the error bars are so enormous that we really don't know which program is stronger. Customers, on the other hand, will make a leap not connected with reality and jump to conclusions not based upon facts from these measures. Therefore, while I doubt very much if they are surprised or even concerned about results in any public experiment (unless a serious problem turns up -- but I have never seen one) I suspect that they are concerned about public perception of these events. The event and public perception of the event are related but separate. There may be some feelings stirring when they see an event turning sour due to a bad run of probability. But these feelings will be in connection with the misunderstandings that will be connected with the event, and not with the event itself. No amount of verbage will educate the great masses of people who view the events. That's because they think they already understand them, and even more so, because the underlying mathematics are entirely uninteresting to them. It may seem like a fine distinction, but in reality it is a very large distinction.
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.