I've been using some computer simulation to supplement playtesting for 2 different game ideas I have. Don't worry, I'm not trying to replace playtesting and I'll still be going through the full gamut. I'm just interested in having more tools to understand why some things work and some things don't. (Also, in its current incarnation, one of the ideas will be very challenging to balance, so it's useful to be able to quickly calculate what the theoretical minimum and maximum benefit/outcome for each card is, without having to expect players to find every powerful combo at human speed.)

Anyway, as it stands I'm effectively running Monte Carlo simulations where I manually set the probabilities of the dummy players doing any particular action. I've got plenty of ideas on how to move forward, both with actual AI, and with tools to explore more specific questions (for example, one game is a deckbuilder, so I'd like to be able to ask the simulator to try to find the earliest time a particular card could be played. I can "ask" this already, but even running thousands of games per second, the decision tree is too large to get reliable data about this specific topic). But before I dive in, I'd like to step back and think about game design to guide my choices.

So my questions are:

• Is anyone else using this kind of simulation in their prototyping? I had assumed it must be pretty common, but I didn't find much on a cursory search. Maybe I don't have the right search terms...

• Do "interesting" or "fun" decisions have any common features visible in a game's decision tree? For example, if choice A is universally better than choice B, obviously you're not in a very compelling situation. A lot of interesting situations I've brainstormed about are those where you have different kinds of gambles: perhaps choice A relies on a small set of conditions, but has a big payout, choice B is similar but on a disjoint set of conditions, and choice C is much more guaranteed but low payout. But the 'magic' isn't just in that description alone, because I can find plenty of bland terrible games that have these kinds of decisions.