The importance of "Dynamism" in Komodo

[lkaufman]

I was curious what does this mean:

The other new option, "Dynamism", gives the user full control over the style of Komodo; a low setting makes it prioritize static advantages like material and pawn structure, while a high setting prioritizes short-term ("dynamic") advantages like mobility and threats to the king. Although the default (100) is believed to be best for actual play, a lower value around 80 seems to produce more realistic evaluations in most positions so may be better for opening analysis.

I decided to use testsuites of different character to check for this setting. STS 1-15 - mostly positional in character, and Arasan 18 - mostly tactical.

STS: 5s/move

Code: Select all

Dynamism  Solved

 80        1346            
100        1367    ------
120        1341

The optimum is at the value of Dynamism 100.

Arasan: 5s/move

Code: Select all

Dynamism  Solved
100         126
140         155  
180         159
190         174    ------
200         171

The optimum is the value of Dynamism 190. The improvement compared to the default number of solved positions is huge, in fact we will see that combined with "King Safety", "Dynamism = 190" makes a tactical monster of Komodo.

So, in tactical problems, high values of Dynamism are favored. Maybe Larry's value of 80 is applicable to very quiet, positional problems?

Then, I also tested several hard tactical suites, and it is possible that Komodo 9.42 Dynamism=190; King Safety=160 (or similar values, try them) is the new King of tactics, maybe at least on a par with Houdini 4 Tactical, way above SF Matefinder. The testsuites were Arasan 18, HardTalkchess2015, DieRybka with about 5s/move.

Komodo D 190 KS 160 beats Houdini 4 Tactical in new Arasan 19 beta suite:
5s/move, 4 cores:
K: 178/250
H: 172/250

My impression is that by varying "Dynamism" from say 60 to 190, Komodo became both the strongest positional engine and the strongest tactical engine. Great feature!

I tried out your /d 190 KS 160 values in actual play (vs default) at 12" + .12", and after 272 games it is down by an astonishing 449 elo! I guess this shows that tactical problem sets have almost no correlation with playing strength.

[shrapnel]

I guess this shows that tactical problem sets have almost no correlation with playing strength.

Disappointing.
Looks like that idea fizzled out fast.

[Laskos]

I guess this shows that tactical problem sets have almost no correlation with playing strength.

Disappointing.
Looks like that idea fizzled out fast.

Not yet, besides that, I was talking of abilities to solve tactical suites. Not yet because I observed that there is a large domain of values of D and KS for which Komodo is exceptional at these suites. It's hard to see what is the best combination for tactics and gameplay, but I seem to have some more moderate values.

[Laskos]

I tried out your /d 190 KS 160 values in actual play (vs default) at 12" + .12", and after 272 games it is down by an astonishing 449 elo! I guess this shows that tactical problem sets have almost no correlation with playing strength.

I tried to find some more moderate tactical sweet points, and I got D=150 KS=140 as performing on tactical testsuites as well as Houdini 4 Tactical (previously the best engine by far in these suites). But the strength goes indeed down very much again, it is about 250 ELO points weaker than default, and 100 ELO points weaker than Houdini 4 Tactical. Still pretty impressive, as the search was untouched, and I think Houdini Tactical has a tweaked search too.

[shrapnel]

It's hard to see what is the best combination for tactics and gameplay, but I seem to have some more moderate values.

Hmm... I suppose you are hinting that there may be some in-between values (between 100 and 150 for Dynamism and King Safety) at which Komodo will perform at optimum, that is, good results for Tactical Suites concomitant with no loss ( or maybe even a gain) in ELO in Gameplay ?
Hmm...lots of experiments and testing required then...

[AdminX]

Interesting watching the side by side.

[Nordlandia]

Ted Summers: Cool!

Anyone with some experience with Table Memory and the optimal settings.

Is 512 MB allocated as table memory too much for 2015 Laptop?

Time control is 180+30 (equivalent tcec superfinal)

If i remember correctly, in the previous versions of Komodo (8 and early 9s)

1. 64 (default) suited for blitz
2. 128 suited for rapid
3. 256 for 60m +/- for each side
4. 512 Larry mentioned 512 for good hardware and long time controls.
5. 1024 for TCEC equivalents.

[Laskos]

Very nice, Ted. Almost different engines, different eval, PV, depth., even NPS. Pity these settings, which seem to be excellent for tactical testsuites, are lowering the strength by hundreds of ELOs. But I guess, with the full range of eval and search tweaks, a strong tactical engine wouldn't be hard to build for Mark and Larry.

[Laskos]

It's hard to see what is the best combination for tactics and gameplay, but I seem to have some more moderate values.

Hmm... I suppose you are hinting that there may be some in-between values (between 100 and 150 for Dynamism and King Safety) at which Komodo will perform at optimum, that is, good results for Tactical Suites concomitant with no loss ( or maybe even a gain) in ELO in Gameplay ?
Hmm...lots of experiments and testing required then...

No, I think for gameplay the default values are close to optimal. I tried some tamer settings, but the decrease in gameplay strength is pretty drastic if keeping the tactical edge.

[mjlef]

Ted Summers: Cool!

Anyone with some experience with Table Memory and the optimal settings.

Is 512 MB allocated as table memory too much for 2015 Laptop?

Time control is 180+30 (equivalent tcec superfinal)

If i remember correctly, in the previous versions of Komodo (8 and early 9s)

1. 64 (default) suited for blitz
2. 128 suited for rapid
3. 256 for 60m +/- for each side
4. 512 Larry mentioned 512 for good hardware and long time controls.
5. 1024 for TCEC equivalents.

Each machine is different, and to some degree, it will depend on memory speed access (which has not kept up with CPU speeds). I suggest just running a say 3 minute search one a position, noting the nps. Then change the Table Memory setting, reload the engine (to clear the tables) then trying again. And keep whatever value give highest nps. As long as you are not exceeding the available physical memory of your machine, the difference will be small.

Mark