First question I'd be asking is, "where does all the money go?"
How much of each entry goes to the prize pool? Is it a fixed amount per entry (really bad), a set percentage of each entry (still sorta bad), a set percentage of all entries (better), or something else?
This is sorta like a progressive-entry fee into a bracket backgammon tournament, where for every level you skip, your entry fee doubles (which is sound mathematically, given full fields and no extra byes). Except the correlation between blinds doubling and/or stack sizes halving doesn't hold the same mathematical validity as in a bracket tournament.
Truth be told, those 200BB you receive in L4 do not hold the same value as the 200BB received in L1, nor is the numerical difference in value anywhere close to linear as suggested by the buy-in table. With an extremely conservative and slow-increasing blind schedule (or with passive nitty players), those later huge stack buy-ins are worth much more than what they are charging. But with an aggressive fast-moving blind schedule (or aggressive swing-for-the-fence players), fewer players will remain five or six levels in, and those players will have much larger stacks relative to the new buy-ins - making them worth less. And as Mental noted above, the physical number of players entering at each level will also skew the true value of late entries.
Interesting concept, but I think it's inherently flawed. The proper way to do something like this would be to base it not on the number of BB at the current blind level, but base it on the current average stack size at a given blind level. Limiting buy-ins to the size of the average stack also protects those players who have earned their large stacks with earlier play, and also helps protects the short-stacks who remain.
