One again @biophil writes a spectacular article! I have a question on one point. You state that you believe that even the worst ICO's increase the total market capitalization. I see a situation where this may not be true.
Although mostly an academic point, if Alice and the rest of the ICO participants buy Magic with Ether they already own the price does not double unless more people enter the market with new USD and purchase ether.
Until the further purchase of ether occurs everyone's Magic coins are essentially worthless. That, or the price of ether must decrease in an amount equal to the rise in the price of Magic. In this situation, the market cap does not change.
Thoughts?
Right, the prices do not double, but the total amount that's accounted for does due to the newly-minted ICO tokens. We know the magiccoins aren't worthless because people just traded valuable Ether for them.
Correct, but what I am saying is that if we have a million ether worth a million dollars to every one magic coin at the release of the ICO, and we take that one ether and trade it for the magic coin the magic coin is now worth 1 dollar, but the total value of ether takes a hit and the one million ether are worth $999,999. or 99.9 cents each. So although it appears to have created two new coins, in reality we have just lowered the price of ether a miniscule amount which is essentially invisible. Once people buy more Ether using USD the price of ether re-normalizes to a dollar and then and only then does the market cap increase.
But why would contributing that Ether to the ICO reduce the market cap of Ether?
Just because the Ether is momentarily out of circulation doesn't mean it's not counted in the market cap. Or am I misunderstanding something about what you're saying?