Correlation or Causation? Interceptions and the Playoffs.
SI.com recently ran an article about the ‘interception ladder’. It looked back at playoff games since 1970, and found an interesting statistic. With each interception a team throws (accidentally throwing to the guys in the wrong jersey) the chance that the team won dropped.
Correlation or Causation? Interceptions and the Playoffs
Now, this is a classic case of confusing correlation with causation. If this data truly was a cause and effect relationship, meaning interceptions caused losses, fixing the problem would be simple, and I could one of the best NFL coaches of all time. In fact, you’d never even need to practice for that gameplan.
If you never threw the ball, you could win nearly four out of five times.
Now, anyone who knows much about football will tell you that you can’t win games like that. What that means is that the interception stat is correlated to winning, but may not necessarily cause winning. For example, a better team will often force the other team into mistakes-the interceptions are the effect of the disparity in skill. Or a team that is losing will make riskier throws to try to score quickly.
The Lean lesson is to read into the data and look for the real cause and effect relationships. Remember: Correlation doesn’t necessarily mean causation.
Request from readers: I’m going to put together a football and Lean page. I’ve used it as a backdrop for making Lean points a few times. I’ve noticed a few other Lean bloggers doing the same recently, such as Mark Graban’s post on value-added time in a football game. If you know of any other Lean posts with (American) football as the backdrop, please let me know in the comments section below. Thanks!