Author: P. Massie
Date: 16:46:50 05/15/02
Go up one level in this thread
I tried that. It doesn't actually change what's in the book. It does greatly reduce the variety of play, so that only the very high probability moves will be played. It also greatly increases the learning function, so that as many games are played the program will (theoretically) learn what to play and what not to play. So these changes should focus Hiarcs8 on a relatively narrow repertoire that it works hard on improving. These factors should significantly improve results over time. Nonetheless, I personally still consider the book great for the player looking for variety, and terrible for the comp-comp competitions. Paul On May 15, 2002 at 17:47:46, Mike J Fitch wrote: > >May i make a suggestion,have you OPTIMISED the settings in book options,you will >find it in TOOLS in the menu.Check the tournament box,and click optimise,maybe >this will stop Hiarcs from playing these second rate moves.Also have you tried >to adjust the amount of ram.I'm running a slow 667.8 celeron with 64 mb ram.When >i first installed Hiarcs8 i tried to maximize the ram which it says was 36mb but >i noticed that my kn/s was very slow/low and my harddrive was working >overtime,so i started adjusting it and checking my Hiarcsmarks,and i found that >at 24mb everything was working great. >Just my 2 cents. > > > > > > > >On May 15, 2002 at 16:50:01, P. Massie wrote: > >>I believe the last two columns in the opening book display represent the >>probability of the move being played, assuming the "Fritz" column (for learning) >>is 0. My Fritz column is 0 - I'm using the book exactly as distributed. It is >>under the Hiarcs8 GUI, and I'm using it to play against myself, rather than with >>computers. >> >>In the Tchigorin line mentioned previously, 2 ... c5 has a 77% probability, so >>it will usually be played. But 2 ... Nc6 has a 9.2% probability, so it will >>show up occasionally. >> >>Also, in the Ruy Lopez line mentioned (1 e4 e5 2 Nf3 Nc6 3 Bb5 a6) the move 4 >>Bc4 is given a 17% probability of being played. >> >>This is the book as distributed, so my conclusion is that this book is designed >>for a wide variety of play against human buyers, rather than being designed for >>optimal strength against humans or computers. I would speculate Hiarcs8 would >>drop a LOT of ELO if using this book against equivalent opposition. >> >>Paul
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.