Welcome to this week’s episode of Hot Takedown, our podcast where the hot sports takes of the week meet the numbers that prove them right or tear them down. On this week’s show (Aug. 25, 2015), another edition of “Stat School” — in which our resident expert Neil Paine teaches us the basics of statistical evaluation in a given sport. Last time in Stat School, we discussed the different metrics of hitting performance in baseball. This time, Neil leads a tour of three ways to evaluate pitchers, each of increasing complexity: wins, ERA and FIP.
Stream the episode by clicking the play button, or subscribe using one of the podcast clients we’ve linked to above. Below are some syllabus notes from Professor Paine to help you follow along.
Stat one: wins
The number of team wins the pitcher is credited with. This isn’t just the number of games a team wins when a pitcher starts — to get the W, he must also pitch at least five innings and be the team’s final pitcher before the half-inning in which the team takes the lead for the last time. Wins aren’t a great stat because the above definition is arbitrary and because they’re so dependent on a pitcher’s teammates. For instance, the offense has to score runs to win games, no matter how well he pitches.
Stat two: ERA
Earned run average is the number of “earned runs” a pitcher allows per nine innings. This is a more sophisticated stat than wins because ERA isolates the pitcher’s job — preventing the other team from scoring. But it’s still flawed. The definition of an “earned” run is based in part on how the official scorer decides to award errors, which aren’t a very good stat for judging defensive support anyway. So while ERA is an improvement on W-L record, it can’t fully zero in on a pitcher’s true skill because it can’t account for the effects of defense (not controlled by the pitcher), luck (not controlled by the pitcher), scorer decisions (not controlled by the pitcher), and even the sequence in which pitchers give up hits or strand base runners (essentially — you guessed it — not controlled by the pitcher).
Stat three: FIP
Fielding independent pitching, part of a family of metrics known as “defense-independent pitching statistics,” or DIPS. DIPS was invented by Voros McCracken, who had the amazing insight (along with an amazing name) that pitchers essentially have little to no control over how often balls in play105 turn into hits against them (as measured by batting average on balls in play, aka BABIP). This insight was revelatory at the time and has largely held up as true since — with some exceptions and softening over the years. As a result, a metric like FIP — which uses only a pitcher’s rates of strikeouts, walks and home runs allowed as its inputs — does a better job of predicting future ERA than ERA itself, because it gives the pitcher neither credit nor blame for the combination of defense and luck that make up his BABIP allowed.
Bonus: There are even efforts to go beyond FIP with “deserved run average.” The inventors of DRA were on Hot Takedown recently — you can listen to that interview here.
Congratulations on graduating from Stat School! Your certificate is in the mail, along with your crushing student loan bill.
Video: Why pitching is so hard to measure
If you’re a fan of our podcasts, be sure to subscribe on Apple Podcasts and leave a rating/review. That helps spread the word to other listeners. And get in touch by email, on Twitter or in the comments. Tell us what you think, send us hot takes to discuss and tell us why we’re wrong.