[Squirrelloid]

Ok, lets talk about the things that do get skewed by how things are currently set up.

(1) you're double-whammying everyone with reduced gold and high upkeep. Ie, not only do they pay upkeep to account for the mage they theoretically buy right away, they also get less gold per turn than they would normally (theoretically to account for mages they aren't buying?)

(2) Many nations are gold intensive rather than resource intensive - giving bonus resources only skews the game (and most importantly the start) in favor of nations which are gold efficient. Some of these nations also benefit heavily from your good/med/bad divisions and upkeep modifiers. For example, not only is arcocephale a good researcher who saves 25% on upkeep costs, they also have many gold-efficient units. I'd have to say they are overpowered with these settings because everything favors them.

(3) The mod premise takes away some element of choice in the pacing of your 'research', and then makes you pay for it. Many nations do not hire a mage round 1 and sometimes not even round 2. Many of those might rely on an awake research pretender for their early research. Those nations are not paying to buy and keep a mage in those early rounds, and can put that money to other essential things (like their army). Some of those may buy a nominal 'mage', but he's really the military leader. (One of Tir Na Og's first commanders is going to be a Sidhe Lord for leading an expansion army - a commander who also has mage capability and can provide research while not leading an army). Some of these nations really can't afford to be forced to lose money this way, since they need to buy those commanders anyway.

I know Tir Na Og felt somewhat crippled just at 75% gold even without the upkeep money for the researchers factored in when i ran some tests. That's ~300 gold/turn from the home province (before dominion effects), which will be -100 upkeep - upkeep for other units. Ow. They can't even afford to buy a Sidhe Lord in a given turn, much less a Lord and some units for him to command.

And not every nation rushes to get labs up. Certainly not in the first 5 turns, say. In fact, while i've certainly started fortresses as early as turn 5, i don't think i've built a lab before turn ~8.

(4) The low gold plus crippling upkeep forces choice of good scales for any nation that's not gold-efficient, and possibly even for those that are. I don't think the proposed conditions are even playable without Order 3, and taking any hits on cash generation at all is probably inadvisable (which means Death, Sloth, and Heat/Cold (unless favored) are all likely bad and possibly crippling decisions).

This has a ripple effect, because needing good scales means either pretenders will have little magical ability, be asleep/imprisoned, or both. That dramatically cuts the playable space for pretender gods to a rather small portion (low magic awake SCs and imprisoned bless chasses really), which dramatically favors some nations at the expense of others.

(5) The good/med/bad researcher distinction totally ignores most of the existing research advantages and disadvantages nations have. Nations with sacred researchers should be able to research cheaper than those that don't. Ie, arcocephale should have more expensive research than a nation like kailasa, who in a normal game will spend far less overall because their upkeep is dramatically lower. Instead, you seem to have reversed that particular example, and i'm sure that's not the only inconsistency.

You've also enforced a unified tempo to how research grows, something which isn't true for all nations (some nations have efficient researchers but you buy small increments so its hard to get lots of research up early, others have researchers which individually provide a lot of RP, but are less efficient per RP). I'm not convinced your good/med/bad distinction really covers these distinctions well, and of course those distinctions are part of how the game is currently balanced. Now, where the externalities are isn't as immediately apparent to me as the sacred/non-sacred researcher distinction is, but i bet there are clear winners and losers in your proposal if someone more experienced than i thought it through. So I can't tell you what the problems are here, but i'm sure they exist.

Edit:
(6) We might also consider how long do you have to pay an exorbitant price to pay off the non-payed cost of the mage who would have provided those RPs. In the hypothetical 8RP mage for 250 gold example, assuming he costs exactly 25 gold in upkeep, then it only takes 10 turns at 50 gold/8RP to pay off his initial cost as well. Except in your model you keep paying that extra 25 gold indefinitely. And heaven help you if your nation's research mages are sacred - now you're paying 4x as much in upkeep as you should be, and will have payed off the initial cost of the mage in ~7 rounds.

Further, your progression grows faster than linearly, while the cost of RPs grows linearly. (Each mage costs as much as the mage before him, assuming same mage type).