I'll answer and accept your terms of mutual dismissal without prejudice.
You
create a Mexican standoff when you are under threat, you can find a non-threatened ancestor to your threator in the threat graph that you can threaten and your other options are worse than creating the standoff would yield. This requires such things as evaluating the value of a threat graph and finding where you can add a threat that would most benefit yourself (most likely by most un-benefiting your threator.)
The mechanics behind all this could be pretty complex but should be able to grow emergently. For each other actor you need a threator and threatee perception value (how much you perceive the threat to and from that actor to be). You need to be able to estimate the chances of the threat resulting in a death of the threatee and whether the threatee will be able to act before death takes place. You need to be able to decide whether to threaten another actor. You need to be able to re-evaluate how the threat graph would look after your threat, including what result your threat is likely to do to the threat graph.
The one big problem, however, is that this is exponential, but it's not that bad since larger than 20-creature fights are rare in adventure mode.
Resolving a Mexican standoff happens when someone's perception changes their action. For example, some actors will have short fuses -- being threatened increases their desire to hurt the person they threaten. Some actors who are initially stupid will start thinking out more than one move ahead, as they more accurately calculate the situation they see the chance of them losing getting very high and choose to back off. Some will be cowardly -- being threatened makes it more likely for them to see backing off as the better option.
Bluffing is a modifier on perception. In general you want to increase the perception that you are a lethal threat when you are threatening someone and decrease the perception that you are a lethal threat when backing off, so your skill in bluffing would change that. If you want to add /causing/ distractions then you need to add an estimate as to distractability of actors. Self-sacrifice would require you valuing someone else's life more highly than your own but should be able to just modify the "cost/benefit" number that comes out of the back end of whatever calculation is there. Bare stupidity would be not being able to look many moves ahead or discounting the value of your own survival or an enemy's lethality.
If you want to open up another thread on this topic when we both have more time PM me and I'll keep track of it. I'm a big fan of behavioral modelling and wish more games did it more realistically. DF could end up being the most advanced behavioral modelling sim if it wants to be.
Author
Topic: Dwarf Fortress 0.34.11 Released (Read 245744 times)