Author: Dann Corbit
Date: 10:29:09 01/06/99
Go up one level in this thread
On January 06, 1999 at 02:05:50, Will Singleton wrote: >On January 05, 1999 at 22:22:37, Dann Corbit wrote: > >>On January 05, 1999 at 22:01:26, Matthew Herman wrote: >> >>>How exactly are we supposed to use this? >>>Is it sorted by eco code? I have the 4000 or so positins in a database now but >>>wondering what to do with it. >>If you just want to look at the positions and see the suggested moves, you can >>try Manfred Rosenboom's excellent EPD viewers: >>http://www.rebel.nl/epd2diag.htm >> >>If you want to use the data, you could (for instance) use the data as a simple >>opening book. Do a lookup based on the EPD string and return the ce value and >>pv, for instance. >> > >Dann, > >Great job on this stuff, and I look forward to the future results. I download >the primary data, and I'm going to use it initially to supplement my book. That >is, when I can select from several alternatives, I'll use this to weed out bad >choices. Also, when out of book (mine is rather thin), I'll switch over to >yours, and also make use of the pv (stuff it into the hash table). > >However, I don't know if the ce is valid for the entire pv, or loses validity at >the end. What do you think? The ce is good only for the current position. After all, they were generated by chess engines. Those same engines will change their minds when they get to the next position. However, our next phase of the project, Project Heartwood, will produce a very usable book, as it will contain the most frequently played positions analyzed in detail. Once simulated annealing has been run on the entire database, the data will become exponentially stronger.
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.