I was thinking about something similar. Chess engines are an interesting kind of software in the sense that they are a form of "competitive programming" which makes standard open source licenses a less than perfect fit in some cases. There are strong arguments for the code being publicly available for inspection because:
It's the best way to remove ambiguity about derivatives etc.
It makes discussing technical aspects easier which I as an author am interested in.
Competitions may require it.
However if the engine is meant to "compete" you may not really want to follow an open source development model. In my DIY engine I probably wouldn't merge pull requests from others. In fact ideally I wouldn't even want other people to create them or otherwise create public forks. Since that creates ambiguity if I ever implement similar features.