Nathan Loofbourrow writes, regarding the on-line game:

You could impose an annoyance factor on the distribution of initial resources (a week's wait, perhaps), or on unlikely transfers (you can't give away the resource except in exchange for some other form of goods... an Objectivist's paradise, perhaps?)

Some kinds of play-by-mail games have had a similar situation, but they have generally not faced the problem in this form because they charge money to enter. This puts a cap on how many entries a person is willing to make. With a large number of participants, controlling two or three players instead of one does not increase the average person's chance of winning enough to make it worthwhile. This does suggest an alternative form of "is-a-person" credentialling, though. Rather than trying to verify identity at a distance, one could simply have a "he paid me $10" credential. You would give these out (probably just one per customer rather than multiple ones) as blinded signatures for anybody who sent you the cash. These could be substitute is-probably-a-person credentials on the theory that most people wouldn't be able to waste a lot of money purchasing a great many of these. OTOH, it's not clear that anyone would be willing to pay this much for a credential unless it had some real, tangible benefit (otherwise it serves as an "I'm a sucker" credential), and if the benefits are great enough perhaps it would be worthwhile to buy multiples. Hal