[Squirrelloid]

i would argue that the relevant factor is not total number of troops bought, its the ratio of mages to non-mages purchased that matters. Ie, the thing of importance here is scale independent - it doesn't matter if there's 100 troops or 1000, what matters if 10% or 40% of cash resources are spent on mages.

I would argue that ideally mages would be made to represent a smaller fraction of your total purchases. Making gold scarce does the opposite, because it makes the number of commander buys relatively larger to available cash.

The two ways to decrease mage purchases relative to unit purchases are as follows:
(1) increase available cash (Ie, the larger the players cash pool is relative to his #commander buys, the more money he'll probably spend on units)
Issue: build more fortresses to buy more mages
(2) make all mages capitol only (hard limit of one/turn)
Issue: nations with better mages are big winners here

supplemental to 1, making fortresses/labs more expensive will stop mage production facility spamming. Actually, making fortresses cheaper and labs much more expensive will make unit recruitment centers cheaper while making mage recruitment centers more expensive.

Say 1/2 price fortresses, triple price labs (and no nation gets cheap labs). At 400/fortress and 1500/lab, players might think twice before building multiple sites for recruiting mages.

(Note: such a price change would make the 'lab burns down' event unacceptable - since you're already running a script, perhaps a script that checks to see if the event occurred, and reruns the turn if it did).