I have a hard time seeing Trump going below 450 unless the states he's currently carrying by 15-20 points have his support all geographically confined. If he's winning a state by that much, he's likely to win every Congressional district and thus sweep the delegate count.
Just did a mockup of the Democratic Super Tuesday, which is much easier -- it's mostly proportional, though apparently a number of the states also do the "by Congressional district" thing as well.
Couple of interesting notes:
1. Clinton could actually be shut out of any delegates in Vermont. Almost all the states have a 15% minimum threshold, and she's polling around 11.4% there. Not a big deal, only 26 delegates, but still somewhat embarrassing.
2. Even though it will be a bloodbath in terms of Clinton likely winning every state other than Vermont (and possibly Massachusetts), the overall delegate disparity won't be quite as brutal as it might look at first glance. I have Clinton taking 627 delegates, to Sanders' 380. Still, that leaves a total gap of 273 (not counting superdelegates). She would be less than 50% of the way to the nomination (if you include her superdelegates), but not by much. And this is a simplistic model -- doesn't look at the potential effect of states awarding some delegates by congressional district, which will likely pad Clinton's total and cut into Sanders' total.
EDIT:
The beautiful irony of the GOP race is that all the various hurdles that the GOP establishment has put in place to try and stymie insurgent challengers (minimum thresholds, winner-take-all and winner-take-most contests, awarding delegates by Congressional districts) are the exact thing playing against them in trying to stop Trump (and to a lesser extent Cruz). Wonder if we'll see some rule changes after this cycle.