Many sports “statistics” seem to ignore causality altogether. The most recent example that comes to mind for me was during one of the recent ACLS games, when an announcer observed that all of the Red Sox’s RBIs had been by left handed batters – so they were really getting it done tonight. This ignores the fact that, save for the case of a home run, the necessary condition for an RBI to occur is for a runner to be on base already. Surely some of the runner’s on base must been right handed hitters – so if no right handers were on base, there would be no left handed RBIs; a whole chain of events had to occur for the RBI to happen.
What of the idea of the “winning run” (touchdown/shot/field goal etc), or its converse the losing play? If your team is up by 10, and you make an error thats lets s run score, that’s no big deal. If you’re tied, it is. Did Bill Buckner lose the world series in 1986? Not by himself he didn’t. Did the Immaculate Reception win that game for the Steelers? Not if they were down by more than 7 points at the time.
This probably all depends on your ideas about causality. If you make two more touchdowns in the first half, then you wouldn’t need the last second heroics, but then your opponent would have played differently from that point on as well – a whole tree of parallel universes of game outcomes.
I love that there’s a list of famous hail mary plays on wikipedia.