The problem with Meditation amulet isn't that you can theoretically get a lot of mana out of it and recover your initial investment in a few turns, it is that it takes up your full action every turn to do it.
Right. I addressed this in my last sentence. Full actions are a bigger deal to builds that
can't replace them with spawnpoints or familiars. I don't disagree with the risk, but I do on how ginormous that risk is.
First, say you're a Druid that uses Vine Tree, Druid's Leaf Ring, Taratree, and Mana Flower.... Now, say you're a Druid who uses 2 trees and the Meditation amulet (and probably also Taratree)....
You raise a valid point about evaluating cards in a vacuum, worrying less about the rate of return on an investment, and
more about what to do with it.
But this scenario raises issues for me.
First, it would be better to have an actual apples to apples comparison. You're kinda moving the goalposts subbing in the amulet and another tree for the two most inexpensive components. Under the same conditions, Ring/Tree/Butterfly/Amulet is more fair, which I have at 6 rounds, -20 mana, -4 quick, and -2 full for +4 profit and +5 full/quick (assuming 2 full plants or less from druid). If you spam enough fulls from vine tree, you can counteract the downside enough to make the extra 4 mana worth it IMO, and if you milk more, that's gravy.
But secondly, none of this matters anyway. Any build that spends two turns farming up will likely scoop to rush no matter what cards you use (unless you're Wizard). This is nothing more than a classic example of what we call, over in the MtG community,
magicchristmasland, which is short for "this works only if your opponent does nothing". For example, I would never count on Butterfly, aka Flameblast bait, going past a round. And the same
I-see-you-playing-cards-that-do-nothing-until-late-game-holy-crap-I-go-kill-you-now arguement against amulet applies here tenfold. You're way better off going spawnpoint/amulet, spawnpoint/mana rock, or spawnpoint/one finger discount and being done with it.