I am quite sure that if you would be in charge of picking 50 opening lines for a TCEC super final or a LTC match between the top engines, we would get a lot of boring draws and a draw rate north of 85%-90%.
To keep a LTC match between the top 3 engines interesting (i.e. more decisive results and more lively chess), you HAVE TO feed them imbalanced lines and something to play with. Feed them the best opening moves and you will get A LOT of (boring) draws.
Komodo 10.4 vs Houdini 5.01 Pro 40+4Min
Moderators: hgm, Rebel, chrisw
-
- Posts: 6052
- Joined: Tue Jun 12, 2012 12:41 pm
Re: Komodo 10.4 vs Houdini 5.01 Pro 40+4Min
was that your book, I did not understand that?Jeroen wrote:I am quite sure that if you would be in charge of picking 50 opening lines for a TCEC super final or a LTC match between the top engines, we would get a lot of boring draws and a draw rate north of 85%-90%.
To keep a LTC match between the top 3 engines interesting (i.e. more decisive results and more lively chess), you HAVE TO feed them imbalanced lines and something to play with. Feed them the best opening moves and you will get A LOT of (boring) draws.
you have been measuring long biassed book vs short unbiassed one, of course the lb will score more wins.
a short biassed book will score more wins than long biassed one(same eval, just use 2-moves book with +70cps edge)
but why do we need biass at all?
engines of equal strength will score many draws, that is about natural.
Komodo vs Andscacs will score very few draws, even less against wekaer engines. Engine X(300 elo stronger than Komodo) will win 95% of the games, no matter the book.