Yeah, sure, as I said this is not going to be the finish version, however I do want to get a feeling for it before going further, so I appreciate any comments.
However it would be better to have them in the mod thread
http://forum.shrapnelgames.com/showthread.php?t=43679
If any of the mods can move them, I'd be grateful.
Regarding national disadvantages
I didn't intend to copy normal nation research progression.
I tried to stick to some guidelines while making the list
a) Good/Bad research depending on how progressive/traditionalistic and cultured the nations is
Arco is thematically about the rise of philosophy and science so good res
Niefel is revisionistic (awake the old giants that tyrannized us at the dawn of time) - so bad res
b) Cost of research is based on the status/accessibility of the mages
Kailasa has a caste system with the mages in the highest caste - you don't expect them to work for a peasant's wage
In Ulm mages are distrusted yet plentyful because of the widespread superstition. A shaman could just a wise man or woman brewing potions (oh god, this is RPG talk, isn't it?) or improving their tribesmens' weapons without demanding much except a place by the fire and a beer.
Of course this is subject to interpretation (and I've used that to boost or penalize some nations were I felt more free to interpret) and the list isn't set in stone. It's more a sketch with which I intended to playtest things out.
If it favors some nations, so what? Vanilla isn't excactly balanced, too, and the winners are not generally feared as uber rush nations. Some of the loosers (Atlantis, Yomi, Rlyeh, Tir, Niefel) trouble me, too, and if anything turns out to be out of balance I'll try to fix that, although I'd rather give it a try before deciding.
About upkeep issues
Well, lets suppose I took a lower upkeep (say 1000) for the late game and higher gold settings. Let's say you're getting 6000 gold = 4500 gold netto, give or take a few, so you are recruiting 10 mages and 200 troops each turn. I want to get away from mage and troop spamming and this isn't really it. I don't think it really undermines bad scale strategies, in fact I've been always thinking the opposite way - the higher the gold the more good scales pay of and vice versa.
I can see your point about early problems, though, and have updated it to 60 gp upkeep.