In recent election years, so many candidates have dropped out early after poor showings in either Iowa or New Hampshire, that shortly thereafter the nominees were effectively chosen and the remaining primaries were just pro forma events. That, of course, is the reason states have been leapfrogging their primaries. And, of course, no nominee has actually been chosen at the convention since (I think) McGovern in 1972.
But this year, with so many states having primaries both early and on the same day (Feb. 5, when 22 states decide), and with the candidates not having been able to campaign much in many of those, there's a possibility that candidates will split their wins across the states that no-one walks out with a convincing lead for the nomination. There's even a remote chance that no-one will get a majority of delegates, and so the candidate will actually be chosen at the convention.
I wouldn't bet much money on that, but it's far more likely this year than in any year since I've been casting votes (that is, since '84). If so, it would be the only good outcome of this ridiculously early primary season. There's a real danger that we'll know both party's nominees by mid-February, and will have to listen to them campaign for nearly 10 months, in which case everyone in the country will be thoroughly sick of the new president before he/she is even inaugurated. One lesson presidents learn is that they can't give major speeches on TV too often, or the public tunes them out, and a campaign requires them to appear on TV, demanding the public's attention, non-stop. Uncertainty up until late August about which two people we'll be choosing between in November would be ideal, and helpful to the future president.
But, in general, the "tsunami Tuesday" is a terrible thing. Coming so soon after Iowa and New Hampshire, where most candidates have been spending most of their time, and encompassing states from Alabama to Oregon, there is no way the candidates can campaign effectively in more than a few of these states. This is why a national primary, which some people advocate, would be a terrible idea--it would just result in candidates having to stage national campaigns at a time when they don't have the funds to do so, and would make it impossible for them to actually meet people.
I've become convinced by the advocates of rotating regional primaries. A regional primary would allow candidates to focus their money and efforts on a limited area, say 5 or 6 states, rather than trying to cover the whole country. Underfunded candidates would be less likely to be completely shut out of the process so early, and the public would still have some chance to meet the candidates. Unfortunately, most analysts seem to think we're not at crisis point yet--maybe we'll only hit the crisis point in 2012--so we're not likely to fix the problem before the next presidential campaign heats up.
There is, of course, a logical stopping point for how early the states can hold their primaries--the day after the previous midterm election. Given that we came close this time to having the first primaries nearly a full year before the general election, that outrageous outcome no longer seems as unlikely as it should be.