I'm pretty sure the pyramid is a rhombus, making a cube at the nadir of power level also gives you a very narrow set of cards to play with. Both extremes rely on power outliers.
I guess that's true theoretically. But at some point, the "weak power outliers" like maybe
Zephyr Spirit, have such awful rates that they're indistinguishable in terms of play pattern from Magic cards whose rates are OK but they just suck to play with, like
Eager Cadet. If power level is defined by the size of the set of cards that create an indistinguishable play experience, I don't think those play patterns are meaningfully different. (Put somewhat differently: I will have just as much fun playing a cube made of curatedly weak cards as randomly weak cards, which is to say, not much fun.)
Some of the advantages presented for higher power don't make a lot of sense to me. You can model it after competitive formats for a lot of free data points and you get a more "powerful" play patterns, some gameplay experiences just by nature rely on stronger cards (or custom magic rules) to exist. Having fewer cards to choose from for each set release is only a positive if you have no self-control, it is objectively just a constraint with no upside, and the second point I'm not entirely convinced is a point the way it's written. Invalidating more expensive threats is presented as the advantage, but that's not a net-positive, it means you layer your cube with trap options, and a lower curve is also not inherently correlated with a higher power level. Strong cards will generally be more expensive regardless of whether or not they cost a lot of mana to cast.
I agree with
@Nanonox that there are subjective upsides, including a vastly decreased cognitive load while making updates.
And regarding the "trap" options -- as the designer, you just cut the cards that no longer hang at that power level. Like, you've outclassed Hill Giant by playing Watchwolf, so cut the Hill Giants, and you move up in the power landscape.
Maybe it's not 1-to-1 correlation, but I don't know many high power formats with high mana curves. You could contrive a low-power low-mana format, but I don't think that's the usual way cubes evolve.
And I agree that the best-in-class cards across Magic will generally be more expensive. But because of the way Constructed works, the 2nd best Tarmogoyf variant sees very little play, but EDH does play the 2nd best beeg Eldrazi; therefore, it is cheaper to build with Territorial Kavus and Watchwolves than with Ulamogs and Kozileks. I glossed over these nuances to write cogently, but thanks for highlighting these ideas.
