The advice from Yorsh sounds good to me. And we also ran short on time, even without a puzzle game! So I think it's pretty universal haha. Another way to implement the AI could have been to just completely randomize moves? Although I don't know how long that would have taken either. In any case, this seems like one of those simple to learn hard to master games, and I'm super impressed that you were able to come up with it and even get it as implemented as you did within the time frame. This is one of the most complete feeling games I've played this year, I definitely think this warrants more work, if you wanted to keep developing it! Really good job!

Cool base for a strategy game, that looks really interesting !

I think aiming for a smart AI in a game jam is too ambitious. Even if somehow it looks doable, there are a lot of ways to fuck it up and have a random bug/unwanted behavior make you miss the deadline. Oftentimes the same ideas could be presented as a puzzle game.

In your case, with the same code you could have made a few puzzle levels, which are much more game jam friendly since :

•  You can change the number of levels you do based on the time you have.
•  You can do quick and shitty levels if you have no time.
• The time it takes to design a level is fairly stable, you can't really create massive bugs and stuff.

Now I'm telling you that, but I was also really short on time on my puzzle game .So take it with a grain of salt :)

Nice idea, could actually make a good physical board game too.