Quick Battle Design Observations

[kevinkin]

Made a small battle with a custom map using the overlay feature. It is a small regular battle with company (-) vs company (-).There is one set-up zone for each side and one AI plan for each side. Each AI plan has two groups - one for the infantry and one separate for the support weapons like MGs and mortars.

As a battle the units follow the AI plans very well.

As a quick battle the new units assigned do not. Remember, in a Quick Battle the OOB is removed to support new OOBs as selected during Quick Battle set=up. Here the units all behaved under the AI plan Group 3. The computer seems to disregard Group 2 (Group 1 not in play). This was observed 3 times with infantry only and once for armor only.

I asked myself, how would you program to allocate units to more than one AI group and if you could would the assignments make tactical sense?

So maybe what I observed is normal and we just have to design with this in mind.

Any thoughts are welcome.

Kevin

PS: are there any types of QBs in short supply. Battle size, environment, type of units etc. ? I would like to start making the needed ones for posting.

[womble]

My previous fiddling with this (in CMBN, so it may be different in the OstFront) suggests that the "top level formations" are assigned, round-robin, to the different orders groups. So, if you have an infantry company (from an infantry battalion) and a tank platoon (from a tank battalion), one will be in one orders group and the other will be in another. If you cull the three infantry platoons from three different battalions, they will be spread across more of the groups available. Similarly for the tanks. Though you'd probably rather they were kept coherent in one orders group.

It's only really practicable to start picking the AI's forces out of multiple battalions rather than one though, when you're looking at larger battles. The cost of extra battalions to prune down to a single platoon can be nearly half the cost of a platoon. You could add (even more) points to the AI to compensate, but it starts to strain the credibility of having a points system at all when so much of the budget is going on "administrative overhead".