the more i look at it, the less appalling it becomes. i wouldn't be surprised if it was only a 1 in a hundred thousand type of occurrence.
the losing team played slowdown against the press, which is always a gamble, as you are allowing more time for the press to create steals. they gave up 17 steals and 22 turnovers, which is not outside the realm of reasonable. the #1 team in the country, playing uptempo, pulled in 25 points off turnovers for a +23 margin on points off turnovers. definitely not average, but not that unlikely, either.
also, the winning team plays a bing-like setup with guards carrying the load and shooting lots of 3s. at fast pace no less. that is an extremely volatile strategy, and if it goes well, you win by huge margins. to say it could reasonably (1%ish or better) cause a +20 point swing by itself is not an overstatement, IMO.
finally, the losing team only took 36 shots, making 8. pretty bad, but they were up against a good press defense, playing slowdown (which should theoretically result in lower fg% than normal tempo in general). if 40% was expected, that would be 1.2%. at 45%, it is about .27%. at 40%, leaving 3pt% at 0, that is a little under a 13 point difference. at 45%, it is a little more than 16.
if you take each event as a 1%, and add the differential, it is roughly 1 in a million, and 59 points. the spread was 51 points, and the better team won. so taking that into account, i feel this is only on the order of 1 in a hundred thousand kind of event. there is a lot of fudging in there but i would guess it is within an order of magnitude of the actual value.