[Twan]

I think you may always try to refute common wisdom with theory, but as long it isn't refuted by results of dozen of MP games it just isn't refuted.

I think the base of said common wisdom is games don't finish in turn 20. And if they take good scales people want something still usefull by turn 50 or 80.

Order is good in early and late game, production is not. Growth is as good for income as production in early game, 10 times better in late.

There may be an exception for the few nations having recruitables using magical weapons and with sufficient hp and mr to survive against late game magic (I don't see a lot... LE Atlantis perhaps ?) so heavy national troops can remain usefull, but it's extremely niche.

Drain is bad from early to late game for the 90% not drain immune nations, sloth is only really bad for early expansion with a limited number of ressource intensive nations and if they don't take an awake god. Magic lose power in endgame once research is maxed, but maxing interesting schools / reaching uniques first is such a big advantage a nation who used drain will probably be still weaker than a magic nation 20 or 30 turns after all finished researchs.

But magic has some side effects that may make it more or less interesting, and may even justify in rare cases to take drain (out of researchers quality, some other things may be considered : do your nation usually use many mages and fight long battles (= profits a lot from fatigue reduction) or has the kind of troops making battles short / has better ways to destroy ennemies than mages spaming spells ? do your mages use MR spells ? do your nation use thugs or undeads whose main weakness is against MR spell ?).

Luck is not as good as order in early game (when your empire is small) and not as good as order in late game on big maps (if your empire is big and expanding, you run into the artificial limit on number of events, and the bad ones in recently conquered provinces often replace the good from your luck 3 lands). Luck is anyway better than order for some situations in midgame (if you have about 20 provinces, *all in your dominion and maxed in scales*, so you reliably get 2-3 good events a turn, luck clearly beats order) and is never a bad choice as it's the only gems scale.