It occurs to me that the emphasis on procedural generation of a storied world plays a part in this model design system. It seems that the game deliberately allows extreme events to take place so that you can get for instance a 'country' which has bad tanks or bad machine gunners or what have you. There have been nations in the last century that never got their head around how to do some part of their military well, emphasising a doctrine or idea that just didn't work.
It is also the case that this happened with some repetition, owing in part to the cultural and economic circumstances (projects changing hands completely and being started again) much like you can have happen in the model design system.
I'll admit, it's irritating to plan on using tanks in a game and you can't get a good roll within the timeframe you can afford. Then again, maybe you rolled a country that is just bad with tanks as a concept. Not every time you play are you going to end up with a superpower culture that is good at everything.
Ofcourse, I'm roleplaying a little bit here to explain what are effectively random dice rolls that bear little relation to what else has happened in your game. It would be better if there was a cause for the effect, some kind of explanation for what is going on and can be seen and responded to before you discover "well, can't use tanks this game."
However, like most things, it might have to wait while far more substantial game elements are added like navies and leader stories that use all of those statistics they've been given. In the mean time, perhaps it would just be best to let the player toggle an option to reject a model and redo it (for full time and price) with a bonus added, which is cumulative if you reject a model over and over up until a certain cap (you don't want unlimited bonus or you could get just a perfect design).