| |
BGonline.org Forums
Best way to improve your BG?
Posted By: Chuck Bower
Date: Thursday, 27 December 2007, at 6:34 p.m.
I've often been curious about the best way to improve one's game. Here I'm talking about high level play. The intermediate probably does better reading books, searching the internet for articles and matches, etc. But once you've read all the books, then what?
Personally I haven't found that playing matches helps me that much. (My game seems to be stuck for the last several years -- here I judge based upon my SW error rate, which has been pretty steady.) I suppose it may depend a lot on one's memory and mine isn't the best -- far from it. But I think there are better ways, even for players with good memories.
The method I describe is one I've learned from both Neil Kaz and Perry Gartner, both of whom use it extensively. It's easy to implement, but like anything, takes time, and a lot of it. But what do you expect, instant improvement?? The reason this method works today (but not 15 years ago) is that you need a strong, trustworthy bot. Even in 2008 with quad-core processors easily affordable, statistically significant rollouts still take time, and why stop and get out of the mood? You need to rely (for the most part) on the evaluation power of either GNU-bg or SW (or both). [JF is a good opponent for play, but without AtS and cubeful capabilities, plus it's now inferior engine, it just doesn't cut it anymore. I know Frank Berger continues to work hard with BGBlitz and it is probably good enough for this method, although I've never taken the time to download/install so I can't speak from experience.]
The simple key is this -- don't just look at a position and say "Oh, play A is right by xxx" and then go on to the next interesting situation. Instead, move the checkers around, (or change the matchscore or cube position) and see what it takes to make play B better. Fiddle around. If you don't understand how factors (like builder positions, strength of the homeboards, racing equity) affect the result, then you really don't know the position. Because in most cases, even if you have a photographic memory, you'll NEVER reach this exact position again. You'll get close, but then these other factors can change the outcome. How deflating it is to find a position you think you know, then make the play you "learned" was right, only to find that it is wrong!
So, when you get a few consecutive hours, try this out and see what you think.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.