The fat lady hasn't sung yet.

The "inevitability" argument comes up frequently from Dean backers, usually with the addendum that if he comes in first in Iowa and New Hampshire, he'll have the race sewn up. Dean may well come out with the nomination when all is said and done and he has certainly run a very impressive campaign thus far. But given the way that delegates are apportioned this year, annointing a winner before any votes are cast is awfully premature.

CAVEAT: My understanding of the delegate apportionment this year is sketchy at best, so if I've made any methodology errors (which is entirely likely), please let me know.

New Hampshire sends 27 delegates to the Democratic convention. Three or four of these will be superdelegates and not committed on the basis of the primary results. The remaining delegates will be apportioned proportionally to candidates clearing the 15% bar. The latest ARG tracking poll (which, keep in mind, completely missed McCain's victory in 2000) has Dean at 35%, Clark at 20, and Kerry at 10. For argument's sake, let's say that by the primary, each of these candidates picks up an additional 5% from independent and Republican crossover voting (I suspect those will largely go to Clark and Lieberman, but as I said, argument's sake) and last-minute defections from smaller candidates, making the numbers 40, 25, and 15. They would then divide the remaining 23 delegates roughly: Dean 12, Clark 8, Kerry 3. The full national convention will have 4,320 delegates.

In other words, a 40-25-15 Dean win in New Hampshire would give him a lead smaller than 1/10 of 1 percent in the overall race. Not exactly a sew-it-up blowout win, is it?

As Mark Schmitt notes, many voters have only just started paying attention and things can move very quickly from here. With crossover voting, New Hampshire is an extremely difficult primary to handicap - much more so than a caucus state like Iowa. The same goes for South Carolina. The nomination race isn't over; it has barely begun.