Compiling Stockfish on Windows

[ilari]

I have followed Eelco idea..and install GitHub for windows..register myself and make a fork with Stockfish..and get same screen as Eelco when everything was setup..

Okay till here..but when i click on "revert commit" i get a window saying:
"You have unmerged files in your working directory.Resolve them and commit them before trying the revert again."

That's what's called a merge conflict: Git's automatic merge tool could not revert the commit cleanly so you'll have to resolve the conflict manually, ie. edit the source code of the files that contain conflicts. The Net is full of documentation, guides and tutorials on how to resolve Git conflicts.

If you seriously want to maintain your own fork of Stockfish that's different from the official version then I'm afraid you'll need to learn at least the fundamentals of Git and C++ sooner or later. Forking things just ain't easy, and keeping forks in sync with upstream is definitely not easy. I'm sorry if this sounds discouraging.

[jpqy]

Thanks Ilari..

The easiest way would be that the Stockfish team add again these 2 settings!
Problem solved for me..

JP.

[mcostalba]

Thanks Ilari..

The easiest way would be that the Stockfish team add again these 2 settings!
Problem solved for me..

JP.

Hi,

I am sorry this github experience is creating a lot of problems for you. Actually the two options were removed because although the naming was nice, the two options were very technical and users keep insisting in asking questions / documentations/ manuals etc about them and at the end any easy explanation was not satisfactory and the real explanation was too technical and really out of the reach of a casual user. So I decided to remove them and magically a lot of pressure on the "questions" support subforum of Stockfish site disappeared. And I am glad of it.

Given that there is really no reason for a user to tweak them, apart from a catchy name that was stimulating user curiosity, I have removed...and I am sorry but I don't think to revert them back.

Thanks
Marco

[jpqy]

Marco i understand your decision..

And i would not ask for it ,if i don't get better results with these 2 settings..
therefore i don't give up so easy on this github experience..
Also i can't use it..but like with your steps i can make a compile now!

So..if it's possible to add these two settings in my own compiles and it's not to difficult to do myself then i need just a guide like , how to add these two settings in a compile.
Once i have these steps in a text-file..i do myself then!

After 4 January we get a lot new compiles ,non off them are better then the last one i have with these 2 settings more.
It will take time to pass SF121227 in my list (with changed settings).

Kind regards,
JP.

[gladius]

Marco i understand your decision..

And i would not ask for it ,if i don't get better results with these 2 settings..
therefore i don't give up so easy on this github experience..
Also i can't use it..but like with your steps i can make a compile now!

So..if it's possible to add these two settings in my own compiles and it's not to difficult to do myself then i need just a guide like , how to add these two settings in a compile.
Once i have these steps in a text-file..i do myself then!

After 4 January we get a lot new compiles ,non off them are better then the last one i have with these 2 settings more.
It will take time to pass SF121227 in my list (with changed settings).

Kind regards,
JP.

What are your new values? We could give them a run on Fishtest.

[mcostalba]

And i would not ask for it ,if i don't get better results with these 2 settings..

Ok, here is the bet !

If you are able to find two parameters of cowardice and aggressiveness that are better than default I will revert the commit back in master !

[jpqy]

That sounds a fair deal!

I will PM you Marco and we can talk further by e-mail for example..

JP.

[Eelco de Groot]

Re-add 'Cowardice' and 'Aggressiveness' UCI options

Congratulations Jean-Paul!

According to Jean-Paul this setup should be stronger
than default.

And SPRT test seems to confirm it:

At fast TC 15"+0.05
ELO: 3.33 +-2.7 (95%) LOS: 99.2%
Total: 25866 W: 5461 L: 5213 D: 15192

At longer TC 60"+0.05
ELO: 7.27 +-5.0 (95%) LOS: 99.8%
Total: 6544 W: 1212 L: 1075 D: 4257

bench: 5473339

Signed-off-by: Marco Costalba

Good result!

[jpqy]

Thanks Eelco..and also for your explenations

It's because Marco i start to learn make a compile,because i wanted these settings back..needed a extra code-line to get them back,but was from older source..so i had to prove that my settings where better..and Marco keep his word!
I had to pass 3 steps in the Stockfish Testframe ,and it did..and now in the new compiles you see them back!

Regression test after amazing JP patch on UCI parameters tweaks.

master vs sf_2.3.1_base diff

ELO: 18.66 +-3.0 (95%) LOS: 100.0%
Total: 17503 W: 3538 L: 2599 D: 11366

JP.

[Houdini]

Well done!