One thing that is lost in the furor over the Barry/Fisher noncall is whether the Spurs made the right move by going for the stop. Statistically, I think they would have been better off trapping Kobe to try to make him give it up to Gasol (who had just missed two FTs) or Odom (shooting well, but historically shaky).
If you look at the percentages, let's say that they have a 70% chance of making each FT, given the pressure of the situation and their relative skill. That gives you a 50-50 shot that they would have missed at least one of two, giving you ~20 seconds to design a play to get a 3.
Even if the shooter makes both, you're down 4, but you still have plenty of time to get a good look at a 3. If you make that, all of a sudden you're down only 1, so when you foul, regardless what the guy does, it's down to a one-possession margin. There's probably something like a 30% chance that they could make a three, given that Barry was shooting well and Manu had just made one.
Compare that to what happened: there was only a ~4 second differential, and you can expect some time to elapse battling to secure the rebound. That means that best case, playing for the stop, you get the ball back down 2 with under 4 seconds left. Given the way games are officiated in the last few seconds, I would guess that there's less than a 10% chance that you get a score in that situation.
Even Elliott's Memorial Day Miracle consumed more than 2 seconds, and Stacey Augmon could have just pushed him out of bounds, probably with no call, if there had been less time. Same thing with the Manu to Horry play in Game 5 in 2005; with less than 4 seconds, Manu probably would have been forced to throw up a prayer instead of dishing back to Horry.
Looking at the cumulative probabilities:
50% that the Laker shooter misses 1 FT * 30% chance Spurs hit a 3 = 15% chance of tying or winning.
50% that the Laker makes both * 30% chance Spurs hit a 3 * 50% chance Laker misses 1 FT * 40% chance Spurs score a 2 to tie = 3% chance
50% that the Laker makes both * 30% chance Spurs hit a 3 * 50% chance Laker makes both * 30% chance Spurs make a 3 to tie = 2% chance
Add those scenarios up, and you get about a 20% success rate for fouling early. Not great. But compare that to not fouling:
65% chance Lakers miss a shot after draining clock * 75% chance Spurs get rebound * 20% chance Spurs score with less than 4 seconds = 10% chance of tying.
Obviously the figures change if you adjust the probabilistic estimates, but not by that much. Any way you slice it, I think the Spurs chances of pulling the game out, albeit slim, would have been more than twice as good by fouling early in the possession.