fallow vigil
Hello all, I like to introduce and persuade some of you into STYLE programming, consider the introduction at https://rebel7775.wixsite.com/rebel/bocc-vs-eas
Inspired by the idea of Stefan Pohl EAS tool I decided to write my own, the BoCC tool ,this page shows the differences and the similarities. Both are meant to list the most aggressive chess engine. As such a new sort of competition made its entry, STYLE PROGRAMMING. If you are bored with ELO PROGRAMMING or need a temporarily break, you hit the elo wall, if you are fed up with 90% draws, want something new : consider STYLE PROGRAMMING and create an engine that plays fantastic chess, a new challenge.
It is much more simple with NNUE than with HCE programming, just feed the neural net with much king-safety and high mobility positions, increase the scores if needed and test your style engine with the BoCC and EAS tools to measure your progress. A new world will open.
In a nutshell, EAS considers 3 evaluations : sacrifices, short games and bad draws. BoCC has 2, king-attacks and short games, the latter in a different way than EAS.
We maintain 3 rating lists of different elo pools, the strongest of CEGT 3578 elo, the medium of CEGT 3500 elo and an entry elo pool of CEGT 3400 elo, the latter available for engines >= 3100 elo, such as HCE Komodo 14.1 with its Aggressive option and doing well. Meaning each of the top style engines (Patricia, CSTal, Rebel, Velvet) has to play 3 x 15.000 = 45.000 games for a reliable style rating.
Well, there is much more but this will do for the moment. Our current STYLE engines are Chess-System-Tal- 2.1 and (just released) Rebel-Extreme 1.1 which you can find at the above listed page.
My wish, I am pretty some of you can a much better job than me.
normal sage