That would be very cool, do you think anyone can do it? I assumed that was way beyond your capacity as most humans cant do that, and the bots in the current competition aren't even close. It sounds like you should use the real time ladder templates.
You just have to break it down to smaller problems. Out of the top of my head, thing one could try:
1) calculate how safe each superregion is, based on how many borders they have with others
2) calculate a rate for expansibility (how many turns it will take to take next superregion once you have that one)
3) determine importance of each region as chokepoints in expansion strategies
4) simulate different scenarios to estimate best expansion strategy and how to counter it
5) estimate opponent income and expansion strategy
6) simulate min-max what the opponent would likely do next turn and the best way to counter it
with these points alone i think you could already code a decently competitive bot against a regular warlight player. taking it to match elite players is another story, but yes i believe it can be accomplished to a competitive degree.
regardless, my idea would be to keep things open source as much as possible, so that new coders picking up the problem on the next challenge already have the grunt work code available and can focus on improving it in terms of strategy.
