Here at Yanks Go Yard, we like to employ every metric, projection, scouting report, and hot stove rumor available in order to provide informed, evidence-based analysis. While simple stats like batting average and wins are nice for certain purposes, we can make much better evaluations if we use more advanced stats, stats that often carry the label of “sabermetric”.
Yesterday, Jimmy Kraft gave an excellent lesson on the theory behind, and uses for, wRC+, which is the #1 metric I use to evaluate offensive performance. You can find that piece here. Today, I’d like to give you, the wonderful readers of YGY, an introduction to DIPS theory, and by extension, metrics like FIP.
DIPS stands for Defense-Independent Pitching Statistics. That means exactly what it sounds like. DIPS attempt to isolate the events that are independent of defense – namely strikeouts, walks, and home runs. These three outcomes are entirely unaffected by the quality of the defense behind the pitcher.
Sabermetrician Voros McCracken first came up with the idea of DIPS over ten years ago. In one of the most surprising, controversial, and groundbreaking discoveries in the history of sabermetrics, McCracken came to the following conclusion:
There is little if any difference among major-league pitchers in their ability to prevent hits on balls hit in the field of play.
Now this statement is probably a bit stronger than it should be, given more recent research on the subject. However, the main point still stands – pitchers who do exceptionally well or exceptionally poorly on balls in play (minus home runs) tend to be unable to sustain that performance in following years.
FIP, or Fielding Independent Pitching, is the concrete form of DIPS theory – it combines strikeout rate, walk rate, and home run rate, and adjusts it to be on the same scale of ERA, and that’s it. It doesn’t care about singles or doubles or triples, or the order of plays. All it cares about are those three outcomes that are most under the pitcher’s control. For those curious, here is the formula:
FIP = ((13*HR)+(3*(BB+HBP))-(2*K))/IP + [constant that changes year to year]
Before we get into some Yankee-specific examples, keep in mind a few points:
- FIP is a descriptive, not predictive, metric. Its entire purpose is to measure three outcomes that pitchers have the most control over. It is not meant to predict the future, though it does predict future ERA better than ERA itself.
- Fangraphs’ WAR (fWAR) is based on FIP. They do this not because they believe that balls in play don’t matter, but because it’s too hard to tell how much they are the pitcher’s fault, and how much they are the fielder’s fault.
- There are many examples of pitchers consistently overperforming or underperforming their FIP. Mariano Rivera is a prime example of this. However, a very large sample size is needed before we can be confident that doing so is skill and not luck.
Ok, now that we’re done with the theory behind DIPS and FIP, let’s get into some specifics. In looking at the past few Yankees seasons, two interesting cases revealed themselves: Ivan Nova and Freddy Garcia. Both had surprisingly strong seasons in 2011, and both completely fell apart in 2012. If we just look at ERA, it’s difficult to understand why this happened. However, DIPS theory gets us closer to an explanation.
Consider their 2011 and 2012 seasons:
Nova and Garcia both saw a huge spike in their ERA from 2011 to 2012, and while some of that can be attributed to the home run rate of the two pitchers, much of it is explained by the percentage of balls in play that fell for hits (which is measured by BABIP). When we look at the FIP values of the two pitchers, we see that there is much less variation between the two seasons. There was still clearly a decline in production, which we should expect given the ERA jump, but DIPS theory tells us that maybe Nova and Garcia weren’t quite as good as it seemed in 2011, and not quite as bad as it seemed in 2012.
I mentioned Mariano Rivera before, so I want to quickly reiterate that a pitcher sometimes does have sustainable control over balls in play. Over his long career, Mariano Rivera has a 2.21 ERA, but “only” a 2.75 FIP. That’s a huge split, and is largely due to his ability to hit the corners and jam batters, thus leading to weaker contact and more outs on balls in play.
Nevertheless, Mo is an exception. For most pitchers, success or failure on balls in play is just not sustainable long-term. Strikeouts, walks, and to a lesser extent home runs, are the better stats to look at if we want to see a pitcher’s true skill. It’s tough to accept the idea that such a high percentage of plays have such little impact on a pitcher’s value. However, if we want to remove luck and defense from the equation and truly isolate a pitcher’s performance, FIP is one the best tools we have.
All stats and references came from FanGraphs. For more information on DIPS theory, check out their glossary page here.