Interesting idea but I think it could create some problems and so it should be approached cautiously. For one thing it is possible that the now inactive player would not have kept the alliance for very long (possibly not for even for one more turn). Also human players are always free to break both the letter (as it were) and spirit (such as there is) of alliances they make. If AI were not written to do so, we human players would have even more advantage over them then we have already. Sometimes it is in the best interest of a risk player to break an alliance, so AI should not be constrained in that regard. To avoid these potential hazards in implementation of this idea AI would need to be allowed to break their alliances at any time, even immediately if that was in their best interests. But that opens a new can of worms. AIs are (probably) programmed as generally as possible. What one thing constitutes a breach of an alliance? An attack of any kind? An attack on a allied continent? How would an AI determine that a human (or other AI) had broken an alliance between them? Just to be clear I am not entirely against this idea. I only want to encourage more discussion of it before it is implemented (if that ultimately happens).
about 5 years ago
Bots can never have the ability to make alliances, nor keep them as being proposed. Alliances are truly annoying to begin with and clearly take the fun out of playing RISK. I almost always introduce bots in the games I create specifically to keep the game moving quickly and fairly.