Most statistics that you hear on ESPN are stupid because they are so meaninglessly specialized ("Hernandez is only the 6th player in the last 12 years to hit 8 times in his first 5 games". Whatever.) But one that I heard on ESPN a couple of nights ago stuck with me not because it was stupidly complex, but stupidly simple.
The talking heads on Sportscenter were setting up the next game in the NBA finals and trying to make some point about the dramatic significance of game 5. Breathlessly, one of them announced that with the series level at 2-2, "the team that wins game 5 has gone on to win the series 18 times out of 24." Wow, pretty impressive sounding stat, right? That all-important momentum from winning game 5 carries you to victory.
Except that if you think about it for two seconds -- which is about how long it took me -- you realize that there is no game 5 momentum. None whatsoever. Nil. Zilch.
If you assume that both teams go into the last two games with a 50/50 chance of winning each game -- i.e. no "momentum" for the game 5 winner -- 18 series wins out of 24 for the team that wins game 5 is exactly what you'd expect by chance. Consider: the team that is down 2 to 3 needs to win both remaining games. With a 50% chance in each, it has a 25% chance of winning both games and the series, so it should win 6 out of 24. Which is just precisely what has happened historically. Winning game 5, in other words, has conferred no special "momentum", that makes the team that wins it no more likely to win game 6 or 7. Each of the following games is still a toss-up.
In short: the magical "momentum" of winning game 5 is that you lead 3 to 2 and only need one more game to clinch the series. Nothing more, nothing less.