"Testing the Finite State Machine in Saga" Topic

For those of you that read my first two articles on using Finite State Machines (FSM) for miniatures gaming (http://solo-battles.blogspot.com/2012/12/finite-state-machines-and-miniatures.html and link I have added the third (and final) article in the series, where I test out the FSM built for playing a programmed opponent in Saga: link (If you have not read the first two articles, I suggest you read them before the last. If you do not know the Saga rules or concepts, definitely read at least the second article before reading the last.)

One of the lessons learned, both from the game play and from the comments on the blog, is that FSMs are simply one trick in your solo gaming bag of tricks. You cannot seriously game using this only this method; you will spend far more time writing "the program" than you will gaming! (Of course, that is a theory. You might, after enough effort, reach a tipping point where your program is rich enough to cover the majority of contingencies.)

I was concerned that I was using too foreign a subject as an example. Saga, for all its buzz, is still far eclipsed in popularity as say, DBA. (With a new draft of the DBA 3.0 rules, perhaps it is time to pick up my DBA Solo project again…) But the comments so far indicate that the example is understandable. That has always been my complaint about solo gaming systems – not enough concrete examples of their use. I hope this is of some use to you, even if you do not play Saga.

Dale

Very interesting experiment. I scanned through "The Solo Gamer's Bag of Tricks" post of 3/4/13 and wsa interested by the 5 basic decision mechanisms (Just Do It, Logic-Based-Chance-Confirmed, Reaction Based, Rule-Based, State Based). I liked the idea of a meta-system that chooses a decision mode for the opponent at various or random times. I could see some advantages to this (as I'm going to write a bit, my apologies if I'm not up enough on your work to have caught these ideas on your blog).

The meta-system would seem tweakable for historical purposes. For example, native armies in colonial settings might be more prone to reaction-based decisions, while armies from totalitarian regimes might be more rule-based. and state-based decision making sounds congenial to armies using auftragstactik.

It would also seem a suitable way to account for officer quality. Some commanders might be capable enough to choose the most responsive decision option (Just Do It, perhaps, or Logic-Based, Chance Confirmed) while others merely respond to events or go on a sort of fuzzy autopilot. Others might be so hidebound that they will obey a rules-decision method without thought.

Anyway I thought it so I posted it! Thaks for your work.

That is an interesting concept – to use different methods for different personalities. It goes along with my idea of "trading" programmed opponents with other solo gamers, so you can fight a personality different than your own.

This is occupying my drive time. I was thinking, too, that there could be guidelines for applying each type of decision making within a single command structure/personality. Some commanders/armies might go into one or another decision mode depending on factors such as enemy behavior, terrain, or the point in time of a campaign. Say there's a general who's scared witless of elephants charging, but is fine with other situations. Or a commander who's been given freedom of action until the enemy penetrates to striking distance of a certain city, at which point the Supreme Dictator and his War Council issue "not one step back" rule-based orders or something of the kind.

I have started a Programmed Opponents forum (on Nabble, hanging off of my blog) for these kinds of discussions. Although it is at the start of trying to take off, that is where we have started to discuss ideas. All welcome.