TalkChess.com

Komodo 12.1 is planned for release this morning at komodochess.com. While the base program has only one minor bugfix from Komodo 12, perhaps worth two or three elo, the MCTS version has been massively revised and improved. Estimated Elo gain on one, two, or three threads about 100 points, and since the thread limit has been raised from three to twelve the gain grows even larger with each added thread. "MCTS Hash" has been added so users with lots of cores and memory can allow MCTS to run for hours instead of minutes per position before filling up the Hash. Testing groups are encouraged to test the MCTS version, especially on four (or even up to 12) threads, as we believe it will gain more than most normal engines from MP. I'll go out on a limb and predict a CCRL blitz rating of around 3120 and a CEGT blitz rating not far below 3000 on one thread, and at least a hundred more on four. The playing style should be similar to Komodo 12 MCTS, that is to say more like a human than normal engines, and less materialistic.

Since Komodo 12.1 is primarily a fix for various problems in Komodo 12 (mostly in MCTS), it is free not only for subscribers but also for all those who purchased Komodo 12. Further improvements will probably be free only for subscribers. We anticipate further continued rapid progress with MCTS but not very much with the normal version. Perhaps MCTS will catch normal Komodo within a year, but who knows?

Congratulation. Am looking forward for the MCTS one. I have lots of things to test where Komodo fails and I know where.

I just have to ask: are there perhaps plans to enter Komodo MCTS into TCEC 13? My understanding is that TCEC rules would actually permit entering both classic and MCTS versions of Komodo, as the two are fundamentally different, i.e. the MCTS version can not be considered a "derivative" of classic Komodo. It is already very competitive, which is remarkable, so I would really like to see it in action - and I'm certainly not alone.

Branko Radovanovic wrote: ↑Mon Jun 04, 2018 12:09 pm I just have to ask: are there perhaps plans to enter Komodo MCTS into TCEC 13? My understanding is that TCEC rules would actually permit entering both classic and MCTS versions of Komodo, as the two are fundamentally different, i.e. the MCTS version can not be considered a "derivative" of classic Komodo. It is already very competitive, which is remarkable, so I would really like to see it in action - and I'm certainly not alone.

We have not discussed entering multiple versions yet. Some might find it unfair. Back in the past at the World Micro Computer Chess Championship, companies would enter multiple identical chess machines to better the chance of coming in first. But we have noticed Komodo MCTS really beats up on weaker opponents more than regular search Komodo. It could be like a form o super Contempt. But this is not Contempt since Komodo turns contempt off in MCTS mode. So, it might be good to use Komodo in earlier stages (assuming they change back to that system), and maybe switch to regular mode in the higher level stages.

I am not against having both regular Komodo and MCTS Komodo in the same tournament if the other teams agree.

From Larry's post, it seems like the idea is to shift developmental emphasis from traditional search to MCTS. This worries me slightly. Standard Komodo is immensely strong, and runs well on laptops, slower hardware, etc. MCTS may not. I'd really hope that work continues on the standard engine, and with the new linkup with chess.com, perhaps now Komodo can have access to Fishtest-esque hardware needed to chase down new ideas and test them.

mjlef wrote: ↑Mon Jun 04, 2018 3:04 pm We have not discussed entering multiple versions yet. Some might find it unfair. Back in the past at the World Micro Computer Chess Championship, companies would enter multiple identical chess machines to better the chance of coming in first. But we have noticed Komodo MCTS really beats up on weaker opponents more than regular search Komodo. It could be like a form o super Contempt. But this is not Contempt since Komodo turns contempt off in MCTS mode. So, it might be good to use Komodo in earlier stages (assuming they change back to that system), and maybe switch to regular mode in the higher level stages.

I am not against having both regular Komodo and MCTS Komodo in the same tournament if the other teams agree.

That's a good point - I only considered the current situation, where the two versions are not close in strength. If they do become close and both play in Premier Division, then the odds of winning may be indeed unfairly increased.

Maybe the solution is (at least in the new division system) to say "one engine per author per division", meaning that e.g. Komodo MCTS could enter the tournament, but could not be promoted to Premier Division, unless it effectively replaced Komodo classic there. I believe this would satisfy the considerable interest in this engine while remaining fair to all competitors.

Branko Radovanovic wrote: ↑Mon Jun 04, 2018 4:17 pm

mjlef wrote: ↑Mon Jun 04, 2018 3:04 pm We have not discussed entering multiple versions yet. Some might find it unfair. Back in the past at the World Micro Computer Chess Championship, companies would enter multiple identical chess machines to better the chance of coming in first. But we have noticed Komodo MCTS really beats up on weaker opponents more than regular search Komodo. It could be like a form o super Contempt. But this is not Contempt since Komodo turns contempt off in MCTS mode. So, it might be good to use Komodo in earlier stages (assuming they change back to that system), and maybe switch to regular mode in the higher level stages.

I am not against having both regular Komodo and MCTS Komodo in the same tournament if the other teams agree.

That's a good point - I only considered the current situation, where the two versions are not close in strength. If they do become close and both play in Premier Division, then the odds of winning may be indeed unfairly increased.

Maybe the solution is (at least in the new division system) to say "one engine per author per division", meaning that e.g. Komodo MCTS could enter the tournament, but could not be promoted to Premier Division, unless it effectively replaced Komodo classic there. I believe this would satisfy the considerable interest in this engine while remaining fair to all competitors.

I like that solution. Since it would apply to all engines, we wouldn't need an exception to the rules.

schack wrote: ↑Mon Jun 04, 2018 3:33 pm From Larry's post, it seems like the idea is to shift developmental emphasis from traditional search to MCTS. This worries me slightly. Standard Komodo is immensely strong, and runs well on laptops, slower hardware, etc. MCTS may not. I'd really hope that work continues on the standard engine, and with the new linkup with chess.com, perhaps now Komodo can have access to Fishtest-esque hardware needed to chase down new ideas and test them.

MCTS runs fine on my laptop. Anyway we'll continue to try new ideas in standard Komodo, but MCTS ideas get priority over just tweaking parameter values in standard Komodo.

Thank you very much Larry;
we will test this version soon.

As you double the number of physical cores with MCTS do you double "search speed" (or whatever the right term is) or is there a mulithread penalty as with normal engines?