I compile the engines on the android and I often have to conduct a tournament between them to identify possible errors. In Chess For Android, there is a game option between the engines, but many games take a long time, the engines spend a lot of time either on tied positions or on the positions won. Therefore, I spend a lot of time in vain. I already wrote about this on the e-mail of Aart Bik, he said that the award is a good idea, but never did it.
hagtorp wrote:I compile the engines on the android and I often have to conduct a tournament between them to identify possible errors. In Chess For Android, there is a game option between the engines, but many games take a long time, the engines spend a lot of time either on tied positions or on the positions won. Therefore, I spend a lot of time in vain. I already wrote about this on the e-mail of Aart Bik, he said that the award is a good idea, but never did it.
I see. You need adjudication parameters.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
hagtorp wrote:I already wrote about this on the e-mail of Aart Bik, he said that the award is a good idea, but never did it.
I have a very, very long list of feature requests and I although am really motivated to make everyone happy, finding the spare time to make it so is not always easy.....
For my own understanding on proper GUI adjudication, I assume that the "draw" settings of movenumber=35, movecount=5, score=5cp mean that if after move 35 we have 5 consecutive moves where the score is below 5cp, the GUI can adjudicate this as a draw? Do GUI's use their own evaluation for this, or should both engines that are playing agree on the score? I can see issues with either model, so some insights on this would be appreciated.
abik wrote:Do GUI's use their own evaluation for this, or should both engines that are playing agree on the score?
The latter one. Otherwise, the GUI would have to implement a better chess engine than the ones it is adjudicating. It's also easiest for the GUI author.
With the previous version, I had one vague user report where going out of the application and back to it again during a game would make the engine unresponsive. Can you check using a Winboard engine whether this might have been a GUI issue? Could as well have been on my side since it was a dev version, but I've switched over to UCI anyway, so I didn't investigate.
Ras wrote:I had one vague user report where going out of the application and back to it again during a game would make the engine unresponsive. Can you check using a Winboard engine whether this might have been a GUI issue?
Yes, that is a bit too vague
I have no issues running Xboard engines in my GUI. Please make sure the engine is compiled properly for whatever Android version you are using, see this old thread for starters.