Script-heavy AI?

[Megalomatrix]

I was a huge fan of TOAW 2 and 3(Matrix), but one thing bothered me:

It seemed like the AI was heavily script-based, so that formations were not so capable of being used by a central AI to achieve flexible, changing goals, but were more bound by objective tracks.

This was based from my editor/forum knowledge, but: Is that the case for TOAW 4? Or: Do scenarios field AIs capable of using all resources flexibly?

[shunwick]

It's really a function of the TOAW engine's flexibility. When you allow users the freedom to design and create their own forces and maps, then you will have to accept programmed AI. There are techniques you can use to simulate some flexibility of response but it is down to the scenario designer to programme for the expected and the unexpected - triggers, alternative objective tracks, different stances etc. A central AI can only be flexible within the parameters of the objective track it is following. A limited response to nearby objective tracks is the probably the best that can be achieved.

It is possible for the player to modify the programmed opponent on the fly.

How well a scenario creator has programmed the PO varies greatly with both the creator and the scenario he/she is creating. Some maps are easier to programme than others. Some forces are easier to programme than others. Some timelines are easier to programme than others. It is a hugely complex problem. A central AI to handle anything a creator can throw at it is simply not possible with current AI development and we are not likely to see any change to that anytime soon.

Pretty sure that we can expect improvements to the TOAW IV programmed opponent but I wouldn't expect anything truly revolutionary.

Best wishes,
Steve

[76mm]

shunwick, thanks for the interesting response. Are you aware of any tutorials for programming the TOAW AI? It doesn't sound particularly intuitive.

[shunwick]

I don't know of any in-depth tutorials for programming the PO. Such a beast would be very difficult to produce, if it truly goes in-depth, due to the extreme variability of scenarios.

See...

http://www.wargamecollege.net/category/archives/misc-toaw-design/ for an example the sort of things that need to be considered during the design process.

Trial and error is, unfortunately, the only way to go. Start very small and take it slow. Programming the PO for a large and complex scenario is not something that is done lightly or without multiple iterations. Many creators, even well known ones, ignore the PO and just slap a "PBEM Only" sticker on the front when maps, OOBs, and timescales can all be PO unfriendly.

Learning how to programme the PO can be a lot of fun. And can keep you amused for several years...

Best wishes,
Steve

[Megalomatrix]

Thank you, shunwick, for the detailed response. Much appreciated.