Author: Stephen A. Boak
Date: 16:07:50 09/09/00
Go up one level in this thread
I'm in the Los Angeles area. Last weekend I saw CM6000 for $4.99 in a CompUSA store I frequent. That was the full price (no requirement to pay more and submit for a rebate) For that price, CM6000 can't be beat. However it has some good and bad points. The hash table size can be adjusted by altering one .ini file (unlike CM7000 which is limited to only 1 or 2 MB hash table size, unchangeable). It doesn't continuously analyze, move by move, a game you wish to play over on the computer screen (i.e. continuous analysis mode does not exist). It does perform superb, automatic full game analysis (my main use), in two different types of annotations (natural language, and computer) How I use CM6000 (and other programs): I like to analyze all my chess tournament games, some interesting games from my ICC play, and (on occasion) interesting games of friends. For the most interesting games or complicated games, I usually analyze the full game using my three main chess programs/engines: Fritz6a, ECTool Rebel 10 engine (haven't been able to get the Rebel Century engine for ECTool to run properly, as yet) and CM6000. I then compare the suggested best moves/lines/scoring. By comparing the different program move selections and suggested lines of play, I learn a lot regarding: 1. How to play better; i.e. better make move choices (sometimes all 3 programs agree; sometimes 2 only agree; sometimes all 3 programs disagree--there may be several overlooked ways of playing a given position) 2. Why bad moves are bad (especially tactically, but also positionally) (I look at the suggested best moves of the programs and try to understand *why* those moves are considered better than the text moves in the game) 3. Relative strength of the scoring, evaluation and tactical characteristics of the several chess programs. (my findings: I trust Rebel the most for postional understanding; CM6000 or Fritz6a the most for tactics; and Fritz6a for scoring that is positionally and tactically very good to excellent, falling normally somewhere midway between Fritz6a and Rebel 10. What I mean is, Rebel seems to give a not very high evaluation score to certain positions, even when one side has an advantage. CM6000, on the other hand, often 'overscores' the evaluation for that side. Fritz6a scoring typically (but not always) runs midway between Rebel scoring and CM6000 scoring. Occasionally I have graphed the different scores of these three programs, move by move, for entire games, to get an appreciation for the eval and scoring differences. I have learned to never trust a single program for analysis in all positions--each program has strengths and weaknesses for analysis purposes.) With CM6000 I typically use 3, 5, 6 or 10 minutes per move (occasionally the max of 999 sec = approx 16 min) settings for full game 'overnight' analysis. I run the program during the day when I am at work, or at night, when I am sleeping. When I wake or return, I cut & paste the natural language (English) text annotations into a WordPad text file, followed by the analytical scoring & best lines of play annotations. The CM6000 analytical scoring mode gives me the computer evaluation score for a certain move, along with the best suggested move and suggested best following line of play, allowing me to learn better moves and a relative sense as to how much better those are than the actual game moves. The CM6000 natural language annotation format is a bit simplistic for my strength, but it does nicely describe the occasional tactical combinations that were mistakes in the played game, and the better line of play that should have been chosen. It does this in a great way that allows you to see quickly that the line played in the actual game should have lost, say, a Rook for a Bishop, whereas the best line of play would have only lost a Bishop for two pawns. Both lines of play are shown, as well as the description of the gain/loss results (gain x, y, z for a, b, c) of each. My OTB (over the board) rating is approx USCF 1900, but I believe with the use of at least one decent computer program (like CM6000) I can analyze and understand many games (by studying them after completion), as well as a strong expert (2100+) or even master (2200+). The computer analysis shows me the tactical holes or tactical sufficiency of certain lines of play actually played, or that I might have contemplated playing, and the suggested moves improve my own positional sense of what works and what does not work (by giving new moves/scores for suggested best play, while showing scoring of the actual game moves). In fact, I annotate many of my own or my friends' games, using the computer analysis, putting my personal comments (improved by use of the computer for analysis) after the game moves. My comments typically add some positional or strategic human reasoning (not typically shown in English comments by any computer analysis, but perhaps suggested by examination of recommended computer moves), i.e. my own judgements as to what works or doesn't work, and why, augmented by what I learned from examining the computer move selections and scoring and comparing & contrasting that with what I know at my thinking level. All in all this analysis stuff is fun, and improves one's own ability play and analyze (both your own and other's games). For $4.99, CM6000 can be a great bargain for some avid chess players seeking help with game analysis. --Steve
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.