Overall, things are progressing. I have run into a few issues (more later) but, I have worked through the tank models, and I am working on the aircraft models now. I can see that historically these are a bit different, but I will go forward with the general idea. Thus far, the tank models are satisfactory.
There are a few snags, which I am in communication with Vic, and we'll see. They aren't complete showstoppers I am also probably being a bit more complicated. The one problem has to do renaming, which doesn't seem to work quite right within the event that is called to create the model. I could be doing it all wrong, but. So, this could change my plans to have a different naming system. Another is a change that I think would make the model process a bit more flexible, but is likely harder to implement and could have issues with backwards comparability. It would be nice, but 100% essential as of now. It would leave a slight issue with obsolete concepts still being available, but there is also a workaround for that, it is just a little clunky.
At any rate, most of the coding should be done fairly soon, and I have built the subroutines to be adaptable to some possible changes.
I don't think I ever had a capability to upgrade between model families. It should work that if you upgraded or improved a model the previous model will upgrade to that model. The new version will be a bit different in that models will not be upgradeable from Level N to Level N+1. However, when you make a new model of the same type, say Medium Tank, the newest instance will be set to upgrade for the latest model with Level N-1. By, this I mean the model itself isn't upgraded, but you can upgrade the subformations in a unit. That looks to be working, but will need some testing.
As for AI+ and AI++, what happens is that the AI gets big production bonuses, so it will produce lots of equipment and push you hard that way. I haven't looked at just how big the bonuses are, nor have I looked at how to control them. That said, something like this could be a way to make things more competitive.
I have been thinking about the AI quite a bit, and I am not sure if it is something more particular to my mod since the unit structures are a bit different, and the overall code might prefer something different (basically, I am thinking of the basic subformations as companies for all unit types). The AI situation is a bit complicated and depends on a few factors. One might also be if you use the Diplomatic block, which allows you to get set up and form a line. This is harder for the AI, as it doesn't quite know what to do with it. But, unless I miss my mark, what you are probably experiencing is the AI suffering losses of the order 2-2.5:1 (maybe even 3:1). After awhile, these losses add up, and the AI is effectively crippled. These losses largely come from attacks on defensive positions at modest odds. The fact that the AI makes these attacks might also be because I make it fairly aggressive to start a game. After suffering these losses, the AI also is bringing up troops, but they tend to be fairly low in readiness since they are being transferred, etc. I was experimenting, and when I get done with the mod rework get back to, some changes that will help the AI somewhat. One might well be not to use the diplomatic block, and let the AI get its shot at you. But, the others are some changes to stack limits and concentric bonuses to give the AI a little edge to reduce its losses. I am looking at some very simple things, like giving the AI the chance to use three full units from two hexes to attack before overstacking penalties (instead of two) and small bonus for attacks from two hexes. These are probably the most likely types of attacks. The other is to reduce the readiness loss for reassignments and transfers for the AI. I also want to try to code up some "objectives" for the AI. That is try to determine a location for it to target and make that a focus. It will be random, so you won't know.