WAR Primer Volume 3 - Pitcher Runs and Wins

This post is part of a series that includes material originally written for The New York Yankees All-Time All-Stars but had to hit the cutting room floor prior to publication. For other posts in the series you can search for the label "Author's Cut" on this site.

In the first two parts of this series the focus was on position players and how their value is translated into wins. Pitchers are up next as we conclude this primer on WAR.

As mentioned in the previous two installments, if you want to dig even deeper into WAR, Baseball Reference's WAR Explainer is recommended reading.

Runs Surrendered

The total number of runs given up by the pitcher, whether they were deemed earned or not. Down the road this will be compared to the average expected runs surrendered by the pitcher to arrive at a wins total.

Fielding Adjustment

Astute observers probably noticed that unearned runs are included in the pitcher’s total from the outset and cried foul. That’s a valid beef since that includes runs scored due to poor defense, but there is a better way to handle that situation.

At Baseball Reference, they adjust the average expected runs using the Rfield values described in Volume 1 for the team playing behind the pitcher. Obviously, a team full of good fielders would lower the expected runs allowed for a pitcher whereas more runs would be expected when pitching for a team that struggles in the field. This eliminates the subjectivity involved when official scorers determine earned vs. unearned runs.

Leverage Adjustment

This adjustment on average expected runs applies only to relief innings and accounts for the concept of leverage. For instance, a reliever that comes into a game with 2 outs and no runners on in the 4th inning of a 10-0 blowout is in a very low leverage situation. What they do in that situation will usually have a relatively low impact on the game’s result.

On the other hand, if a reliever is called from the bullpen in the 9th inning of a tie game with the bases loaded and no outs, they are in an extremely high leverage situation. The game will almost entirely depend on how the pitcher performs there. Pitchers that appear in many high leverage situations would be expected to allow more runs than those that don’t.

Pitchers Positional Adjustment

The pitcher’s equivalent of the positional adjustment made for a player’s WAA - as described in Volume 2. When it comes to pitching, there is a distinct difference in both style and results between starters and relievers.

In general, relievers tend to surrender runs at a lower rate than starters and therefore the average expected runs for relievers is adjusted to be a bit lower than it is for starters.

Based on research done at Baseball Reference, it has been determined that this dichotomy is a relatively recent phenomenon so this adjustment is only made on innings pitched since 1960.

Mariano Rivera's 56.3 career WAR leads all Yankee pitchers despite ranking 23rd in innings pitched.

Keith Allison/Wikimedia Commons

WAA - Wins Above Average

Number of wins added by the pitcher above that of an average pitcher. Compared to the WAA calculation for position players, the pitcher WAA is fairly straightforward, but they both start with finding the runs above average for the player.

For pitchers, we need only their Runs Surrendered (as described above) and the number of innings they pitched. The number of runs an average pitcher would have surrendered in the same situation is then arrived at by summing the runs scored by each team the pitcher faced, adjusting for the parks in which they played, and weighting them appropriately by the number of innings pitched against those teams. From there, of course, the other adjustments detailed above are made.

Once all adjustments are accounted for, we have the final average expected runs for the pitcher. If you subtract from that the pitcher’s Runs Surrendered, you’re left with the pitcher’s runs above average. The conversion of those runs to WAA is nearly identical what is done for position players - also described in Volume 2. The key difference being that pitchers can only affect the "Runs Allowed" value in the Pythagorean expectation calculation.

WAR - Wins Above Replacement

A single number that represents the wins the pitcher added to the team above what a replacement pitcher would add. The WAR calculation for pitchers relies on the same assumptions for deriving a replacement runs value (Rrep) that the position player WAR does. Overall, the concept is the same as laid out in Volume 2.

At this point, the WAR and WAA metrics may appear to be too similar to be used independently, but the difference between WAR and WAA can lead to some interesting real life results.

Take the example of two Hall of Fame pitchers, Pedro Martinez and Phil Niekro. Over the course of their careers, Niekro’s 97.2 WAR edged Martinez’s 86.2 career total by a decent margin. This is mainly due to the fact that Niekro pitched nearly twice as many career innings as Martinez. However, in terms of WAA, Martinez is the better pitcher by almost the exact same margin with a total of 61.5 to just 50.8 for Niekro.

Phil Niekro presents an interesting edge case highlighting the difference between WAR and WAA.

tradingcarddb.com/Wikimedia Commons/Public Domain

What this reveals is that Niekro was a classic compiler, having many seasons, especially at the beginning and end of his long career, where he was at or below average but better than a replacement player. During those seasons he bulked up his WAR total while not adding any value above league average.

Martinez, on the other hand, spent virtually his entire, relatively brief career as an above average pitcher and was considered the best pitcher in baseball from the late 1990’s into the new century.

Comparing WAR against WAA can help highlight the difference between players that were truly dominant versus those that were pretty good over a very long period of time.