Retuned values after eval quantize patch.

[xoto10]

The last search tune patch was tested before the implementation of #2733 which presumably changed the search characteristics noticeably. Another tuning run was done, see https://tests.stockfishchess.org/tests/view/5ee5b434ca6c451633a9a08c and the updated values passed these tests:

LTC 60+0.6 th 1 :
LLR: 2.97 (-2.94,2.94) {0.25,1.75}
Total: 11176 W: 1499 L: 1304 D: 8373
Ptnml(0-2): 69, 933, 3403, 1100, 83
https://tests.stockfishchess.org/tests/view/5ee6205bca6c451633a9a147

SMP LTC 20+0.2 th 8 :
LLR: 2.93 (-2.94,2.94) {0.25,1.75}
Total: 54032 W: 6126 L: 5826 D: 42080
Ptnml(0-2): 278, 4454, 17280, 4698, 306
https://tests.stockfishchess.org/tests/view/5ee62f25ca6c451633a9a162

Closes #2742

Bench 4957812

[ssj100]

Nice work congratz! For completion, it also passed:
STC 10+0.1 th1:
https://tests.stockfishchess.org/tests/view/5ee62f05ca6c451633a9a15f

[vondele]

@xoto10 thanks!

I think it might be worthwhile to run another long tuning run on top of this patch, to see if we have reached convergence, or if more can be gained.

[xoto10]

Interesting idea. (I was hoping I was finished with these for a while! :)
I could do 200k games, starting from the latest values, how does that sound?

Edit: I'm assuming LTC. Something I've wondered, does the tuning process work with SMP?

[vondele]

yes 200k (or more) from the latest values sounds good.

I've also been thinking about tuning SMP... I think it should work, but I don't think it has ever been tried, so it could trigger some fishtest buggs. It does start to get really expensive quickly.

[xoto10]

Or more?? I'll do 300k then, we have done that kind of length before and the fleet makes that possible (thanks noob!).

[vondele]

Thanks!