Remove QueenInfiltration

[unaiic]

STC https://tests.stockfishchess.org/tests/view/5f2955b1a5abc164f05e4c85
LLR: 2.96 (-2.94,2.94) {-1.50,0.50}
Total: 29216 W: 5560 L: 5416 D: 18240
Ptnml(0-2): 466, 3329, 6902, 3417, 494

LTC https://tests.stockfishchess.org/tests/view/5f299154a5abc164f05e4ca1
LLR: 2.92 (-2.94,2.94) {-1.50,0.50}
Total: 54144 W: 6635 L: 6594 D: 40915
Ptnml(0-2): 372, 4859, 16536, 4966, 339

bench: 4023423

Remove QueenInfiltration from eval.

[gonzalezjo]

Major deja vu from this PR.

[AlexandreMasta]

But why?

This eval was intruduced just a few days ago.

Why do ppl rush in making such removals?

[unaiic]

It was actually introduced more than a month ago, and just as happened to king infiltration, there is nothing wrong with removing it IMO if it passes the tests.

[ddugovic]

The tests pass and I was surprised #2759 passed in the first place (many evaluation terms are gradual but this term is not).

[FauziAkram]

This Bonus works very well the longer the time control is, which means that it scales perfectly well, it would be really a pity to remove it just because it does not scale super well at STC.
I wish we could test it also at VLTC, and I am pretty sure it will never pass the simplification bounds

[ddugovic]

Hm, I wonder if it is possible (after merging this patch) to write a form of it which gradually increases the bonus based upon how safe the queen is.

[gonzalezjo]

Doesn’t hurt to try it at LTC again. Maybe the last LTC wasn’t a fluke and QI just works better at LTC. I don’t think that’s that likely, though.

[mrobaczyk]

Seems like it is possible that a "new-term" patch can pass a test and then simplification of that patch passes too, which can go forever.

[ddugovic]

Yes, and ability to go forever is not a bad thing, especially not when other queen evaluation patches are merged.

Perhaps next time this bonus is developed, instead of using a fixed S( -2, 14) it could somehow scale with how many squares the queen is attacking or something like that, so it is less likely to be simplified in the future.

[vondele]

Thanks