I was the primary tester for Twelve Clans during development. This is @Michael-Hoover's first time working with xml and my first time trying to seriously test a game and make it as good as we can.
@Cernel those are some good observations. I agree with you completely when it comes to a no unit cap version. I was against capping the units when I first started to test the game myself, actually. But the more that we discussed it and tested it out the more that making caps seemed like a good solution to the AI problems. Our issues with the AI may be fixable in other ways. But caps were the best solution at the time given our limited experience. I also agree that more documentation would be helpful. Please continue to give more feedback as you notice things. Any suggestions are appreciated and will be taken into consideration for future updates.
Hopefully, as the game gets more people playing we can take all of everyone's feedback and improve the experience.
Here is some background on why we settled on the way that the units in Twelve Clans are capped and some of the problems that we noticed during testing.
We noticed that without a cap the AI would only ever buy certain units. They loved Ashigaru in the Alpha, when they were the same as the Bushi (this is before the Bushi existed). To solve that we decided to make the Ashigaru not purchasable. This was inspired by the Dragon War Auxiliary units. Then the AI had a new favorite unit the Ronin. We tried re-balancing the unit stats and prices, but they always seemed to buy tons of the same units. This was especially problematic with Ronin, as they leave after one round (meant to represent mercenaries only following coin) In an attempt to fix this we decided to implement caps.
With the caps in place the AI started purchasing more varieties of units and seemed to be more fair and competent. I am not sure that the AI fully understands that the Ronin leave after one round though because they make some questionable moves sometimes. But it is a lot better than it was prior.