Nemeton 1 release

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Stan Arts
Posts: 179
Joined: Fri Feb 14, 2014 10:53 pm
Location: the Netherlands

Nemeton 1.1

Post by Stan Arts »

Nemeton 1.1

http://1drv.ms/1Dc4J7k

What's new.:

A quick and dirty attempt to set the evaluationvalues to something more realistic. Atleast it's less materialistic and a bit more aware of activity.
Lots of modest speedups, cleanups and improvements to the code.
Threaded Winboard input means that it can now ponder, interrupt search etc. inside a GUI.
Overhauled ponder implementation itself. From just filling the hashtable to actually interrupting and continuing the search on a ponderhit.

Stan
User avatar
Graham Banks
Posts: 41473
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Nemeton 1.1

Post by Graham Banks »

Stan Arts wrote:Nemeton 1.1

http://1drv.ms/1Dc4J7k

What's new.:

A quick and dirty attempt to set the evaluationvalues to something more realistic. Atleast it's less materialistic and a bit more aware of activity.
Lots of modest speedups, cleanups and improvements to the code.
Threaded Winboard input means that it can now ponder, interrupt search etc. inside a GUI.
Overhauled ponder implementation itself. From just filling the hashtable to actually interrupting and continuing the search on a ponderhit.

Stan
Thanks Stan.
Roughly how much strength do you think you've gained?
gbanksnz at gmail.com
Stan Arts
Posts: 179
Joined: Fri Feb 14, 2014 10:53 pm
Location: the Netherlands

Re: Nemeton 1.1

Post by Stan Arts »

Graham Banks wrote: Thanks Stan.
Roughly how much strength do you think you've gained?
Me? I casually work on my chess and my Elo improves steadily. I suppose about 50 Elo a year slowly getting closer to my goal of "expert".

Nemeton? Ponder off, nothing too impressive. 50-100 Elo?
Ponder on vs other engines with ponder on, a bit more than that. With draws against big engines in tests here. Not quite sure yet why that is or how to translate it to ponder off play as well..
User avatar
Graham Banks
Posts: 41473
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Nemeton 1.1

Post by Graham Banks »

Stan Arts wrote:
Graham Banks wrote: Thanks Stan.
Roughly how much strength do you think you've gained?
Me? I casually work on my chess and my Elo improves steadily. I suppose about 50 Elo a year slowly getting closer to my goal of "expert".

Nemeton? Ponder off, nothing too impressive. 50-100 Elo?
Ponder on vs other engines with ponder on, a bit more than that. With draws against big engines in tests here. Not quite sure yet why that is or how to translate it to ponder off play as well..
Sorry - I meant Nemeton. :)
Sounds like a nice gain.
gbanksnz at gmail.com
Stan Arts
Posts: 179
Joined: Fri Feb 14, 2014 10:53 pm
Location: the Netherlands

Nemeton 1.2

Post by Stan Arts »

Nemeton 1.2

Bunch of tweaks. In eval it has a penalty for worse side having fewer pieces which at the same time encourages the better side to simplify the position. That seems to be a surprisingly significant improvement. Had something like that in Neurosis too, now it's more refined.

Other than that it really does seem to perform relatively better the longer timecontrols get. That was my original design/intent so that's good..

Next step is to try and go parallel somehow.

http://1drv.ms/1fexsyT
User avatar
Graham Banks
Posts: 41473
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Nemeton 1.2

Post by Graham Banks »

Stan Arts wrote:Nemeton 1.2

Bunch of tweaks. In eval it has a penalty for worse side having fewer pieces which at the same time encourages the better side to simplify the position. That seems to be a surprisingly significant improvement. Had something like that in Neurosis too, now it's more refined.

Other than that it really does seem to perform relatively better the longer timecontrols get. That was my original design/intent so that's good..

Next step is to try and go parallel somehow.

http://1drv.ms/1fexsyT
Thanks Michael.
Roughly how much of a gain do you think you've made in Elo terms?
gbanksnz at gmail.com
Stan Arts
Posts: 179
Joined: Fri Feb 14, 2014 10:53 pm
Location: the Netherlands

Re: Nemeton 1.2

Post by Stan Arts »

Graham Banks wrote: Thanks Michael.
Roughly how much of a gain do you think you've made in Elo terms?
Hey Bert.
The past has shown you guys are better at determining that than me. :)
User avatar
Graham Banks
Posts: 41473
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Nemeton 1.2

Post by Graham Banks »

Stan Arts wrote:
Graham Banks wrote: Thanks Michael.
Roughly how much of a gain do you think you've made in Elo terms?
Hey Bert.
The past has shown you guys are better at determining that than me. :)
No problem John.
Posting during a toilet break in the middle of the night is never a great idea! :lol:
gbanksnz at gmail.com
User avatar
fern
Posts: 8755
Joined: Sun Feb 26, 2006 4:07 pm

Re: Nemeton 1.2

Post by fern »

at the same time encourages the better side to simplify the position...."

For your info, that was a technique invented by Kittinger and implemented in one of his last chess computers. Perhaps it could be even improved if the acceleration of changes take into account what kind of pieces are still on the board and what kind of advantage is the one.

As an example of what I say, if you are advantageous BUT the pieces on the board are just pawns, that accelerated rate of exchange would not be advisable.

Just a thought of an amateur, not a programmer.

Luck

Fern
Stan Arts
Posts: 179
Joined: Fri Feb 14, 2014 10:53 pm
Location: the Netherlands

Re: Nemeton 1.2

Post by Stan Arts »

SzG wrote: It seems to me that Nemeton still can't ply under WinBoard. Here is a PGN WB created during the first stages of a tournament:
It plays under Winboard fine. I do it all the time.

More helpful would be to explain what kind of setup you are using. Because you are not playing Nemeton with it's own book.