Hello,
I am
the developer of ChessCombiV1 and as of today, I’m starting to
create a new version of ChessCombi.
ChessCombi
is just a small project I made at school. At that time, I was really
interested on chess engines and I was thinking on how to make them
better. This idea of combining chess engines really intrigued me.
There are some aggressive, risky chess engines, and on the other hand
there some safe positional chess engines. And I was interested what
would happen if you would combine both engines. I would experiment on
different chess engines, and sometimes laugh at the results. There
are games that some risky moves will be played and when it’s given
back to the positional chess engine, the score would be negative. I
imagine this as the positional engine “shaking its head” while
the risky engine is saying “What?” cluelessly :D.
There
is also an idea to have different chess engines working together on a
move. I visualize this as chess engines meeting and discussing
together. But I soon realize that this is a flawed approach, as
computer resources is limited (CPU time and memory), the capacity of
2 engines working together will be halved compared one engine working
alone.
As I
was stumped by this fact, after some time I lost interest and I
started to get busy with other things with school and work. I would
still play chess from time to time, but I didn’t touch ChessCombi
or other engines during this time.
But
recently, I read that Leela Chess Zero (an AI) has won the TCEC
CUP-2. Leela beats Houdini in the finals and Leela also beats the
current world champion Stockfish in rankings. Link:
http://blog.lczero.org/2019/02/leela-won-tcec-cup.html#more.
This made me excited because I found that the rise of new types of
chess engines. This new chess engines would learn and improve by
itself. This quite different than traditional chess engines that just
calculates the position and the best moves. This continuous
improvement is something that I found fascinating in the new chess
engines. But I am not ready to “jump ship” or “join the
bandwagon” yet!
I
would like to quote GM Yasser Seirawan, and that is chess at its core
is a battle of ideas. As a fan of traditional chess engines, I would
like to see traditional chess engines will still prevail despite the
rise of AI. I am hoping we haven’t reached that limit on how
calculations are done on traditional chess engines. My idea is that
the traditional chess engines can work together to beat Leela! :D
With
the advent of AI in chess like Leela Chess Zero and Alpha Zero, I
read and watched about chess more and more, and I was
shocked/surprised to find out that people are using ChessCombi
despite of its lack of features and being hard to use. I also checked
my previous yahoo mail account and found a couple mails regarding
ChessCombi. These recent events gave me enough motivation to continue
ChessCombi.
Here
is the plan for ChessCombiV4 features:
->
There will be a GUI so that you can configure the chess engines
easily (no more text files :D)
->
There will be a way to set conditions when to switch engines, so it
would be more configurable.
->
There will be a way to check an outlier in the scores. This can be
used as a condition on when to switch chess engines.
->
There will be some way to save the results of evaluation of the
positions of the games. So, in next games, it can be used to
determine which chess engine will play this line.
->
There will be an installer and a portable version of the software so
that it can easily be used by more users. :D
If
you want more features or you want to contact me, please do so on
detectivemark7@gmail.com
(I rarely use my yahoo account anymore :D).
Lastly,
ChessCombi will always be free. I am firm believer that free software
yields to more users, more users leads to more feedback, and this
ultimately leads to better software.
Thanks
for your time and interest,
Mark
Comments
Post a Comment