2022 Blogmas Day 20 – Pitcher WAR

We’ve already addressed the concept behind WAR (Wins Above Replacement) in the abstract and for offensive players in particular. Today, we’ll consider how we quantify runs for pitchers and overview calculating their WAR.

A Quick WAR Overview

WAR stands for Wins Above Replacement. We consider a replacement-level player equivalent to someone earning the league-minimum salary, such as a minor league player who is replacing an injured player, or a veteran player who is typically on the bench and comes in as needed.

To calculate WAR, we first find how many total runs a player is worth compared to the average player in the league; then, we convert those runs into wins by determining the number of runs per win. Finally, we adjust the result so that it compares to replacement level, not average.

At times, I’ve written far more information about a statistic because either I couldn’t find a good resource, or a good resource had some poor explanations that I wanted to elucidate. That’s not the case for pitcher WAR, so I highly recommend reading the FanGraphs WAR for Pitchers library page. It’s absolutely excellent. As such, I won’t be diving into too many details. I’ll only focus on how different sites calculate WAR¹As a reminder, we have fWAR for FanGraphs, and bWAR or rWAR for Baseball Reference, and where our discussion yesterday of Leverage Index enters the conversation.

Pitching Runs

At its core, measuring runs is way simpler for pitchers. We already have several metrics (e.g. ERA and FIP) that tells us the runs pitchers allow; we then compare the chosen metric²It’s a philosophical choice to use ERA, FIP, or something else as your core for WAR. FanGraphs prefers to use FIP, while Baseball Reference uses (very roughly) total runs allowed per nine innings. Read more about Baseball Reference’s choice here, ironically on the FanGraphs site.to the league average, and thus compute how many fewer (or more) runs the pitcher allows per game.

We scale the runs metric as needed so that it’s on the same scale as the MLB average runs allowed per nine innings. This allows us to create a final runs metric that’s on the same “scale” as the metrics we saw for offensive WAR.

Then, instead of using a static number of runs-per-win like we saw on offense — hitters have far less direct impact on the game, while pitchers have a great effect for some innings, and no effect in future innings — we use a dynamic approach that is unique to each pitcher.

Leverage Index for Relievers

Relievers are unique. Some are brought into high-leverage situations and they should be rewarded for their performance in those. In addition, relievers are not replaced in the same way as starters. They typically follow a hierarchy of duties where the next-best pitcher replaces the best pitcher, and a new replacement-level reliever will get slotted into the bottom of the bullpen, put into low-leverage situations.

So, we take reliever WAR and scale it using their average LI³Strictly speaking it’s the average of their average LI, and the league-average LI of 1.00., thus providing high-leverage relievers who do well a bit of a bump in value due to the effect they have in tense, later innings.

Fun Statistics⁴I used Baseball Reference to find these statistics; in general, pitcher bWAR is a slightly more “volatile” and inflated value compared to pitcher fWAR.

One benefit of accumulation statistics like WAR is that they inherently rely on playing time, so I’m not worrying about “qualified” pitchers.

• There are 6 pitchers with 10 or more seasons where they accumulated at least 5 bWAR: Roger Clemens, Bert Blyleven, Randy Johnson, Greg Maddux, Mike Mussina, and Tom Seaver.⁵Max Scherzer has 9, while Justin Verlander has 8. They’re both still active so they could join this list.
• There have been 12 teams that had 3 pitchers who accumulated at least 5 bWAR during the season. Only 5 seasons were before 2000, and the most recent was the 2019 Washington Nationals⁶They won the World Series. rotation that included Patrick Corbin, Max Scherzer, and Stephen Strasburg.
• In 1972, 1992, and 2009, there were 20 pitchers who accumulated at least 5 bWAR. The most pitchers in a single season since 2000 to accumulate at least 7.5 bWAR is 4, in 2018: Jacob deGrom, Kyle Freeland, Aaron Nola, and Max Scherzer.
• We talk about the year of the pitcher, but in both 1971 and 1972 there were 3 pitchers who accumulated at least 10 bWAR. Any guess for which player made the list on both years?⁷You’re wrong. It was Wilbur Wood.

Continue to Day 21 – Fielding Percentage and Range Factor

• 1
  As a reminder, we have fWAR for FanGraphs, and bWAR or rWAR for Baseball Reference
• 2
  It’s a philosophical choice to use ERA, FIP, or something else as your core for WAR. FanGraphs prefers to use FIP, while Baseball Reference uses (very roughly) total runs allowed per nine innings. Read more about Baseball Reference’s choice here, ironically on the FanGraphs site.
• 3
  Strictly speaking it’s the average of their average LI, and the league-average LI of 1.00.
• 4
  I used Baseball Reference to find these statistics; in general, pitcher bWAR is a slightly more “volatile” and inflated value compared to pitcher fWAR.
• 5
  Max Scherzer has 9, while Justin Verlander has 8. They’re both still active so they could join this list.
• 6
  They won the World Series.
• 7
  You’re wrong. It was Wilbur Wood.