Today is an exciting day in SABR world. Fangraphs launched what may look to the average fan like a few new junk stats. To me, these stats represent the biggest advancement since I started blogging in 2010.
For those who really care about Fielding Dependent Pitching, I’ll let Dave Cameron explain in two pieces what it means, what its components are, and how we should (or shouldn’t) use it to evaluate pitchers. For those who would rather just scrape the surface, here’s a quick primer:
For years, the primary difference between fangraphs’ version of WAR and baseball-reference’s version has been the way they allocate run prevention between pitchers and hitters. Fangraphs bases its pitcher WAR on FIP (Fielding Independent Pitching), which assumes that pitchers are only responsible for their strikeouts, walks, and home runs, and that all other outcomes are the result of fielding and randomness. B-R bases WAR on runs allowed, ignoring subjective “errors” and crediting (or debiting) a pitcher for everything that happens when he’s on the mound. Both parties know that the truth lies somewhere in between- pitchers have some amount of control over how often batted balls are turned into outs and whether baserunners score or are stranded, but it’s certainly not complete control.
As a fan, and as a blogger trying to make an informed decision in award voting, I’ve always made my decision based on a healthy dose of fWAR, perhaps balanced by a dash of rWAR. Sure, I’ll look a little deeper into the FIP components- strikeout rate, walk rate, and home run rate- and even the FDP components- BABIP, strand rate- but none of these numbers is park adjusted, and as rate stats, they ignore volume, so while they may tell me something valuable in my effort to evaluate a single pitcher, they’re not all that helpful in comparing pitchers pitching different numbers of innings in different environments. Basically, I’m stuck between two imperfect versions of the same catch-all statistic.
Four new stats came out of fangraphs’ deep dive into FDP. RA9-wins approximates rWAR, eschewing FIP for a version of WAR based on run prevention, essentially answering the question how much better than replacement level was pitcher X’s team at preventing runs when he was on the field?. The difference between RA9-wins and WAR, then is FDP-wins, which answers the question how much of the team’s run prevention was a result of outcomes other than strikeouts, walks, and home runs? FDP-wins is broken into two components: BIP-wins, or how much of the difference between run prevention and true outcomes was based on balls in play turning into singles or extra-base hits? and LOB-wins, which encapsulates all other fielding-dependent outcomes, but primarily represents the pitcher’s ability to strand runners on base, rather than allowing them to score.
In developing FDP and its compenent stats, fangraphs staff considered changing its WAR formula to reflect some percentage of FDP-wins. Without empirical evidence of the relative responsibilities of pitchers and fielders in driving a pitcher’s BIP-wins and LOB-wins, however, they decided not to change the WAR formula at the moment, admitting that their version of pitcher WAR isn’t perfect, but that it’s not only more reflective of a pitcher’s talent level, but also more descriptive of his past value, than a version based on runs allowed.
Cameron concludes, quite reasonably, that BIP-wins are mostly the result of fielding and luck, but that a pitcher with a particular batted ball profile (perhaps a knuckleballer) might have some control over these outcomes. On the other hand, LOB-wins might say something about the pitcher’s ability to control the running game or to avoid giving up home runs with runners on base. Neither, though, is singularly reflective of one party’s value.
In essence, these new stats have taken BABIP and strand rates, adjusted them for park effects, and multiplied them by innings pitched to convert them from rate stats to value stats. This is a boon for those of us looking to evaluate pitchers in terms of their Cy Young candidacies.
A few weeks ago, I named Justin Verlander at the top of my AL Cy Young ballot, with Felix Hernandez third. King Felix has since spun off a perfect game and another shutout, while Verlander got knocked around by the Royals last night. They’re now neck-and-neck for the award, Verlander leading in fWAR, 5.6 to 5.5. FDP gives us a little more context for Hernandez’s superiority in preventing runs (7.2 RA-wins to 5.8).
As you can see here, Verlander has actually benefitted from a lower BABIP, to the tune of .4 additional wins, but Hernandez has been two full wins better in stranding runners (1.1 to 0.9). This suggests that much of the difference in their relative abilities to prevent runs has actually been based on Hernandez’s ability to keep runners from scoring, rather than on his pitching in front of a far superior defense (and the Washington Generals have a defense far superior to the Tigers). There are some complications here, such as errors having more of an effect on LOB than BIP, since they don’t count as singles, but it certainly speaks well of Hernandez’s abilities, and I’m ready to hand my vote (if the season ended today, of course), to the king.
Felix leads Verlander by 69 points (575 to 506, far ahead of any other AL pitcher) in my True Season Score metric, which aims to bridge run prevention and FIP, so I may have come to the same conclusion anyway, but there’s something comforting about access to the components that make up the difference between the two theories behind WAR. TSS tells me that Felix’s advantage comes primarily in the form of preventing home runs, but without a park adjustment, I don’t know how much of that is a factor of pitching half his games at cozy Safeco Field. LOB-wins tells me that the Mariners have been two wins better with Felix on the mound than the Tigers with Verlander when fielding independent outcomes and balls in play are removed from consideration. This is supported by Hernandez’s superior strand rate (79.5% to 73.8%).
Park-adjusting some of baseball’s more nebulous stats may not seem like a major breakthrough, but if it means that the best pitcher in the league wins the award he deserves, that’s progress. You may argue that voters will keep giving the award to the guy with the most wins, but it’s clear from recent elections that we’re making progress.
The world is not flat. Polio can be cured. Wins are a stupid way to evaluate a pitcher. And Fielding Dependent Pitching is a combination of outcomes on balls in play and strand rates.
Today was better than yesterday.