Recent ideas (last 4 years) tried in ProDeo that either were a small regression (not more than a loss of 10 elo) or an insecure improvement (not more than a elo gain of 3-4) that perhaps after all have a future in case of more test games.
In the hope it serves some purpose to list them for other authors since I am pretty much done playing 15,000 - 20,000 games to ratify a small change.
http://rebel13.nl/prodeo/ideas.html
-----------
Will do the search part later, less interesting anyway.
Ideas
Moderators: hgm, Rebel, chrisw
-
- Posts: 7221
- Joined: Mon May 27, 2013 10:31 am
Re: Ideas
Monday I saw a blitz game KRP -KR where pawn was an h6 pawn and rook was on h8. Only way to keep it a draw was to attack the pawn horizontally from the sixth row which cuts off the opponent king and only in case of h6-h7, Rh6 had to be played.
-
- Posts: 895
- Joined: Mon Jan 15, 2007 11:23 am
- Location: Warsza
Re: Ideas
I have added a small fixed bonus for bishops standing next to each other to Rodent about two months ago. It was a small plus.
Pawel Koziol
http://www.pkoziol.cal24.pl/rodent/rodent.htm
http://www.pkoziol.cal24.pl/rodent/rodent.htm
-
- Posts: 433
- Joined: Fri Jan 16, 2015 4:02 pm
Re: Ideas
Ed~, thank you for sharing your work.
I also have a few more ideas on the wiki page: https://bitbucket.org/zurichess/zurichess/wiki/Home
So far I haven't figured out a way to find (potential) useful evaluation features. Random forests https://www.stat.berkeley.edu/~breiman/ ... htm#varimp have some metrics to compute which variables are most important, but you still need some amount of feature engineering.
I also have a few more ideas on the wiki page: https://bitbucket.org/zurichess/zurichess/wiki/Home
So far I haven't figured out a way to find (potential) useful evaluation features. Random forests https://www.stat.berkeley.edu/~breiman/ ... htm#varimp have some metrics to compute which variables are most important, but you still need some amount of feature engineering.
zurichess - http://www.zurichess.xyz