Fix bug with "excludedMove" for probcut

[MJZ1977]

Bugfix :
"excludedMove" must be skipped in probcut loop too. If it is not skipped, the probcut can exit quickly with wrong return value corresponding to the excluded move.

See (https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/GGithf_VwSU)

STC :
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17130 W: 3747 L: 3617 D: 9766
http://tests.stockfishchess.org/tests/view/5b8460c40ebc5902bdbb999a

LTC :
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 12387 W: 2064 L: 1930 D: 8393
http://tests.stockfishchess.org/tests/view/5b8466f90ebc5902bdbb9a21

To go further : it can be perhaps useful to tune the singular extension search parameters.

Bench: 4694445

[pb00068]

@MJZ1977 Congratulations for the patch. The continuous integration check failed because you missed to terminate your commit message with the bench-number.
Your commit message should be terminate with:
bench: 4694445

[jerrydonaldwatson]

Congratulations, this is a good catch!

[MJZ1977]

I added bench number. Is it OK like this?

[pb00068]

@MJZ1977 Can't see it.
I suggest you to make and push a new commit (containing the bench commit-message) on this branch,
so the continuous-integration test will be triggered again automatically.

[MJZ1977]

Must I make a new branch from the beginning ?

[pb00068]

@MJZ1977 No, not necessary, but app-veyor is still failing.
Just redo as you did, but put the bench into commit-message as separate (last) line, then it should be ok ;-)

[MJZ1977]

Can you please make it, I am not familiar with all these tools.

EDIT: OK I understand now, what you call "message" is called "description" in Github desktop... I changed only the title in 1 line before. :D

[locutus2]

Congrats too! Nice that you found the bug.

The bug was introduced with this simplification 4d64742. Before this the node pruning methods were skipped if we had an excluded move.

Please add to your PR message not the elo value but the complete result like showed on the test page (simply copy paste). See as example the above commit

For your STC this could look like following:

STC : 
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17130 W: 3747 L: 3617 D: 9766
http://tests.stockfishchess.org/tests/view/5b8460c40ebc5902bdbb999a

Also i would recommend @Rocky640's code rewrite.

[MJZ1977]

Considering @Rocky640 proposition, it is not a simple rewrite because the loop will stop after we met the excluded move.
Edit : changed after Rocky Edit

[DU-jdto]

Seems weird to write it this way instead of just if (move != excludedMove && pos.legal(move))

[MJZ1977]

@DU-jdto I agree with you.
I let the maintainers choose the right way to write it. The most important for me is to correct the bug and go ahead. How many time it will need in general to merge with master?

[snicolet]

Merged via 10bb2e6, thanks!

[snicolet]

How many time it will need in general to merge with master?

I suppose you mean "How much time does it need in general to merge with master?"

It depends if the maintainer is in vacation or not, but as a data point I can say that for
each pull request it takes me in general 45 minutes to import the patch, shrink the
multiple parts together, rewrite the code to make the style more coherent with Stockfish
code, correct the English mistakes of the message (and add my own), triple check the
bench, run with the DEBUG flag, push to official master, close the pull request, etc.