Stats: Descriptive and Predictive

With the Pirates back to .500, the argument between the adherents and skeptics of advanced statistics has raised its head again, as exemplified by the first comment in the May 28 game summary:

Some formula somewhere indicates that we lost this game, and fell 8 games below .500

A big part of the problem is a confusion between the two roles of statistics: they describe what has happened, and they are used to predict what will happen. All stats do both - but some are more predictive than others.

All stats are descriptive: they tell you what has happened. The Pirates have won 24 games and lost 24. James McDonald has given up 16 ER in 65.1 innings, for an ERA of 2.20. Pedro Alvarez is batting .215.

Advanced stats are descriptive as well. The Pirates have scored 144 runs and given up 167. James McDonald has given up 59 fly balls, 17 non-intentional walks, 1 HPB, and 63 strikeouts in 65.1 innings pitched, which can be used to calculate an xFIP of 3.27. Pedro Alvarez has 14 singles, 9 doubles, 8 home runs, and 12 walks in 159 plate appearances, which can be used to calculate a wRC+ of 91.

All stats are predictive as well. Or rather, all stats are used predictively - although some do a better job of predicting what will happen than do others.

The Pirates' .500 record is used to predict that they'll end the season around .500 and with a little bit more luck will be in contention for the playoffs. But their pythagorean W/L predicts that they'll end the season 77-85, in fourth place in the NL Central 20 games behind St. Louis.

As SuperBaes referenced in the 5/28 recap, pythag W/L has been a more reliable predictor of a team's W/L record at the end of the season than straight W/L. Yes, teams have overperformed their pythag - but there are far more examples of teams "overperforming" the W/L record you would have gotten by projecting their current W/L at various points in the season.

Pointing out that the Pirates are lucky to be at .500 based on the runs they have scored and given up is not ignoring that they are actually at .500, or dismissing what they've done. The Pirates are at .500 after Memorial Day. This is a good thing. But using this fact to predict that they will end the season with 81 wins and are one big bat away from the playoffs is misusing the statistic called "Won/Loss Record", when there are far more reliable statistics available to predict the season's outcome.

This is a FanPost and does not necessarily reflect the views of the managing editor (Charlie) or SB Nation. FanPosts are written by Bucs Dugout readers.

Log In Sign Up

Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Bucs Dugout

You must be a member of Bucs Dugout to participate.

We have our own Community Guidelines at Bucs Dugout. You should read them.

Join Bucs Dugout

You must be a member of Bucs Dugout to participate.

We have our own Community Guidelines at Bucs Dugout. You should read them.




Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.