PDA

View Full Version : FSM help (no the other FSM)



Rannath
2011-09-11, 09:32 PM
I've become somewhat convinced that a FSM would be pointless for this. I think I'll just maintain a list of actions taken and one readied action. (and a global list of possible actions)

then whenever I get some input I'll see if its valid, if so it'll go into the done list. If you ready an action it gets stored.You can ready an action on your turn and when you've done all possible actions or passed the remainder of your turn its someone else's go.

If a system requires you to do things in order I'll add a list of actions that get duplicated into the list when you perform each action (basically perfoming an action adds all preceeding actions).

NB: as for why I'm doing this, I plan to make an online game/tool for anything with an srd that'll let people make and distribute quests/quest chains/campaigns so they can play table top games online (with integrated voice/video/text chat)

Analytica
2011-09-13, 11:10 AM
My intuition is that this will work for 4E, but not necessarily any other game system. This is because 4E has everything on a clear grid in two dimensions, and a limited list of things you can do. Something like 3.5 with a million spells that can warp the rules of reality, much less feasible in practice. I may be wrong. I and a colleague discussed this when 4E came out, including the option of parsing the PDFs to get all the data automatically.

Rannath
2011-09-13, 08:34 PM
My intuition is that this will work for 4E, but not necessarily any other game system. This is because 4E has everything on a clear grid in two dimensions, and a limited list of things you can do. Something like 3.5 with a million spells that can warp the rules of reality, much less feasible in practice. I may be wrong. I and a colleague discussed this when 4E came out, including the option of parsing the PDFs to get all the data automatically.

urg parsing pdfs is a pain. Especially when some pdfs are just be glorified jpegs.

I'm limiting it to SRD to start.
I'm also generalizing the combat as much as possible. Ex:

Standard
Move
Minor
Free
full-round
etc

I'm also doing all the heavy lifting in scripts so I can change everything on the fly. (For instance the FSM is defined by xml and the actions withing the FSM are defined in scripts)

Also remember: this is a tool to aid a live session, not a stand alone game

Analytica
2011-09-14, 10:50 AM
The real cool thing about state machines, or Hidden Markov Models, though... conceivably you could simulate combats entirely, if you had probability distributions over decisions (which possibly could be optimized for each actor) - then you could determine the expected outcome in nifty ways. This build has an actual 45% chance to defeat Orcus.

Lots of work to do that though, I guess, for pretty useless returns... :smallbiggrin:

Rannath
2011-09-14, 08:02 PM
The real cool thing about state machines, or Hidden Markov Models, though... conceivably you could simulate combats entirely, if you had probability distributions over decisions (which possibly could be optimized for each actor) - then you could determine the expected outcome in nifty ways. This build has an actual 45% chance to defeat Orcus.

Lots of work to do that though, I guess, for pretty useless returns... :smallbiggrin:

I think maybe you're missing the point. The state machine won't be self running. It'll be basically be an engine that'll do all the work that you'd normally have to do at the table top.