There are lots of numbers in a game. But only one is the score.
The pursuit of better data is good. Lots of things can be clues towards where you want to go. But there’s a certain corruption that can happen, and we call it “vanity.”
Imagine a basketball game. Some scientist is in the crowd, and he notices that the cheers seem louder when the home team is scoring more points. So he goes home and pulls up archived footage of hundreds of games and lo and behold, the data fit. It turns out that loud cheers correlate to home team wins strongly enough that you can predict who will be the eventual winner of the game even if all you have is the volume measurement of the crowd’s noise.
This scientist publishes this in an interesting if (it seems to him) inconsequential paper. But then major NBA franchises pick it up and start doing something that seems to make perfect sense to them, but is actually absurd: they start trying to get their crowd to be louder as if it mattered – maybe using extra mascots, t-shirt giveaways, slogans on billboards, etc.
They see the clear science that says “louder cheers = more wins,” but the reality is the reverse – and only the reverse. “Decibels of Cheering” is just a vanity metric. It doesn’t actually affect the outcome, even if it’s measurable and people care about that measure.
It’s easy to get distracted by things you can measure and improve because you can measure and improve them. But “measurable and improvable” doesn’t mean “relevant.” Just because some piece of information is a clue to the real outcome doesn’t mean that changing it artificially can change the outcome itself.
Listen to the cheers all you want, but keep your eye on the scoreboard.