A few years back, I was thinking about PC game AI, and how many games (without particularly complex NPCs) could be replicated in wargames.
Solitaire Wargaming: Designing NPC AI #1
Solitaire Wargaming: Designing NPC AI #2
This was mostly typical theory-crafting, looking at the flow-chart of possible decisions (if : then).
However, more practical folk, among them Nick Caldwell, actually explored this, and created a simple but layered generic system, called SARGE.
This is the link to the document in his blog, Colony 13.
(He's a versatile dude - also kickstarting some minis - 32mm space dinosaurs with jezzails!)
Go to his blog and have a look for yourself. I don't want to deeply dissect something he could probably do on his own blog, but I'd like his idea given more attention (I'm re-bumping it in the google group), as it's both pretty straightforward and has 2 layers of tactics for the AI (unlike any other "solo play" rules I've seen). I'd hate a cool concept to fade into obscurity, as solo/co-op wargaming will, in my opinion, become an increasingly important aspect of the hobby. So I'll give a short summary:
AI focusses on Objectives
There is 3 objective markers which are given a "weighting" - i.e. how important the AI regard them. They "suck" blips/units towards them depending on the value of the target and the proximity of the unit. A bit like gravity - the most important objectives have more "weight." As units/blips arrive, the "weight" decreases (so units will tend to go for important objectives first, but once there is enough units on the objective, excess units will seek secondary and tertiary objectives).
AI "Stances" - Aggressive, Normal, Cautious
To stop units moving predictable in a beeline for objectives, the units are assigned a "stance." The stance is shown by a stance card, which is orientated towards the objective. 2d6 is thrown, and the units direction may be semi-predictably randomised, based on the stance. A aggressive unit will mostly head directly for the objective with only rare deviations, but a cautious unit will move sideways sometimes and even rarely, back away, based on the random-but-predictable 2d6 results.
It's very simple - #1 pick a blip/unit, determine with objective it will go for (based on "weight"); #2 pick the stance card and place it on the unit, orientated to the objective; #3 roll 2d6 and move the direction indicated on the card.
I think there's a lot of room to experiment - i.e. you could set a semi-predictable patrol path by weighting objectives in a sequence, or place the objectives on the enemy table edge for a convoy mission, or even make it so enemy units switch to become the objectives. I've seen other solo systems where NPCs react to the player, whereas this mostly focusses on the objective. I wonder if SARGE could be combined with other systems...
Anyway, this is too cool an idea to sit on a shelf, so go visit his blog; check out the SARGE rules and pester Nick - don't forget to check out the sci fi dinos!
I'm hoping this will inspire others to think about this topic - designing solo AI, not just player rules - as I think solo gaming will be an increasingly important aspect of the hobby in future.
Random Question:What is with all the Swedes visiting this blog? They outnumber the US/UK folk combined by more than 2:1. Is it bots? Is Skynet based in Sweden? Is this because I have been destroying Swedish robots in Generation Zero?<- a great "Dad" paced game, btw. Is "Delta Vector" the Swedish phrase for "funny cat videos" or "hot babes?" I'm very confused.