Expansion of Consciousness (Betsabe, Drofa, Dumb, Jumbo, KnightX, Nawito, OliThink, Paulchen, Shallow, Stash, Supernova)

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

Angle
Posts: 354
Joined: Sat Oct 31, 2020 1:04 am
Full name: Aleksey Glebov

Re: Expansion of Consciousness (Betsabe, Drofa, Dumb, Jumbo, KnightX, Nawito, OliThink, Paulchen, Shallow, Stash, Supern

Post by Angle »

OliverBr wrote: Fri Jan 08, 2021 10:45 am It's possible that OliThink 5.9.2 is somewhat at the end of his journey as I do not see any chance to improve with small changes. Yet.
In order to improve it looks as I will have to make major changes.
This should mean that the existing set of ideas is implemented almost perfectly and weights are tuned almost ideally. It is always a pleasure to test such engines.
Incredibly fast systems miscount incredibly fast.
OliverBr
Posts: 865
Joined: Tue Dec 18, 2007 9:38 pm
Location: Munich, Germany
Full name: Dr. Oliver Brausch

Re: Expansion of Consciousness (Betsabe, Drofa, Dumb, Jumbo, KnightX, Nawito, OliThink, Paulchen, Shallow, Stash, Supern

Post by OliverBr »

Angle wrote: Fri Jan 08, 2021 12:11 pm
OliverBr wrote: Fri Jan 08, 2021 10:45 am It's possible that OliThink 5.9.2 is somewhat at the end of his journey as I do not see any chance to improve with small changes. Yet.
In order to improve it looks as I will have to make major changes.
This should mean that the existing set of ideas is implemented almost perfectly and weights are tuned almost ideally. It is always a pleasure to test such engines.
The implementation is surely not perfect, but I can't find any improvement with minor changes.
There is no tuning in OliThink, it is very small (1500 lines). OliThink has no chess knowledge or data derived from tuning.

Could you say what ELO you derived from the result?
Which is your hardware (if it's written somewhere, just tell me where :)

Thank you again very much, @Angle to let OliThink play in your tournament.
OliThink GitHub: https://github.com/olithink
Nice arcticle about OlIThink: https://www.chessengeria.eu/post/olithink-oldie-goldie
Chess Engine OliThink Homepage: http://brausch.org/home/chess
Angle
Posts: 354
Joined: Sat Oct 31, 2020 1:04 am
Full name: Aleksey Glebov

Re: Expansion of Consciousness (Betsabe, Drofa, Dumb, Jumbo, KnightX, Nawito, OliThink, Paulchen, Shallow, Stash, Supern

Post by Angle »

OliverBr wrote: Fri Jan 08, 2021 4:27 pm Could you say what ELO you derived from the result?
It's rating should be overpriced because most of the opponents were significantly weaker. In my overall rating list OliThink 5.9.2 4 CPU (TC 20s+3s) is 23 ELO ahead of Bright 0.4a 4 CPU. So it's rating should be around 2930-2940 ELO in CCRL Blits rating list.
OliverBr wrote: Fri Jan 08, 2021 4:27 pm Which is your hardware (if it's written somewhere, just tell me where :)
Intel i7-8750H-Quad, Windows 10.
Arena 3.5.1, TC 20s+3s, Ponder off.
Book: libra8.abk.

The processor supports popcnt, pext, avx2, and bmi2 instructions.
Incredibly fast systems miscount incredibly fast.
OliverBr
Posts: 865
Joined: Tue Dec 18, 2007 9:38 pm
Location: Munich, Germany
Full name: Dr. Oliver Brausch

Re: Expansion of Consciousness (Betsabe, Drofa, Dumb, Jumbo, KnightX, Nawito, OliThink, Paulchen, Shallow, Stash, Supern

Post by OliverBr »

Angle wrote: Fri Jan 08, 2021 6:05 pm It's rating should be overpriced because most of the opponents were significantly weaker. In my overall rating list OliThink 5.9.2 4 CPU (TC 20s+3s) is 23 ELO ahead of Bright 0.4a 4 CPU. So it's rating should be around 2930-2940 ELO in CCRL Blits rating list.
Oh, 2930-2940 is impressive, more than I thought. Noteworthy the fact that OliThink uses merely 1 CPU.
OliThink GitHub: https://github.com/olithink
Nice arcticle about OlIThink: https://www.chessengeria.eu/post/olithink-oldie-goldie
Chess Engine OliThink Homepage: http://brausch.org/home/chess
Angle
Posts: 354
Joined: Sat Oct 31, 2020 1:04 am
Full name: Aleksey Glebov

Re: Expansion of Consciousness (Betsabe, Drofa, Dumb, Jumbo, KnightX, Nawito, OliThink, Paulchen, Shallow, Stash, Supern

Post by Angle »

OliverBr wrote: Sat Jan 09, 2021 4:11 pm Oh, 2930-2940 is impressive, more than I thought. Noteworthy the fact that OliThink uses merely 1 CPU.
Ouch, I forgot that OliThink still doesn't support multi-threading! Then the result is even more impressive. But keep in mind that there were only 50 games and only Stash 25 was nearly at the same level as OliThink 5.9.2 (even stronger with 4 CPU). So, the result may be a bit optimistic. Probably, 2900 ELO in CCRL blitz list is a more realistic estimation of OliThink 5.9.2 rating.
Incredibly fast systems miscount incredibly fast.