The “Back to the Basics” series is designed to explore the foundation principals of statistical analysis across the four major American sports. The series will provide readers with an understanding of how teams approach roster construction and why certain decisions are made both on and off the field. Readers will also be directed to additional information sources, such as websites, books, or even magazine articles that could substantially increase their knowledge of the subject at hand.
Fielding Independent Pitching (FIP) is a more accurate representation of pitching performance than the more commonly known stat Earned Run Average (ERA). FIP measures what a pitcher’s ERA would look like if we only factored in measurements that a pitcher can actually control. The origins of this statistic can be found in research published in 2001 by Voros McCracken. In the following paragraphs, we will touch on McCracken’s original DIPS theory, explore FIP, and examine attempted improvements on the statistic.
Defense Independent Pitching Statistics – DIPS
Bear in mind that the following theory was published on January 23, 2001, two years before the release of Michael Lewis’ famed book Moneyball. For many years, it was believed that pitching and fielding were so intertwined that they were impossible to separate in the analysis of a pitcher. In his initial explanation of his DIPS theory, Voros McCracken sought to divide a pitcher’s stat line into what the defense can and cannot affect. By stripping defense from the equation, and only evaluating a pitcher on what he can actually control, Voros believed he would have a more accurate representation of pitchers’ abilities – or at least, one more accurate than ERA.
One of the more prominent findings in the original publication was predicated on what we now know as Batting Average on Balls-In-Play or BABIP. This measures the number of hits a pitcher allows on balls hit in the field of play – meaning that at-bats resulting in home runs, walks, strikeouts, and hit batsmen are removed from the equation [(H-HR)/(AB-K-HR+SF)]. McCracken believed that pitchers have little control over the results of batted balls other than home runs.
“There is little if any difference among major-league pitchers in their ability to prevent hits on balls hit in the field of play.”
It is now widely held that BABIP generally fluctuates on a yearly basis and that pitchers have little to no control over it [some argue that hitters can influence their BABIP levels]. The three main factors of BABIP are: (1) defense; (2) luck; and (3) talent level. (Check out this FanGraphs breakdown for more information on BABIP.)
This analysis by McCracken introduced an entire new wave of statistical analysis in baseball. Today, there are a number of methods available to measure a pitcher’s independent ability. However, the most oft-referenced Defense Independent Pitching Statistics (DIPS) are FIP and xFIP.
Fielding Independent Pitching – FIP
FIP measures a pitcher’s performance after removing the impact of the defense, luck, and sequencing. By removing these factors from the analysis, FIP allows us to isolate the performance of the pitcher and provides greater insight into his future production. For example, a pitcher who has a FIP that is substantially lower than his ERA may be performing adequately, but may be experiencing bad luck or poor defense behind him (resulting in an inflated ERA figure). This is why FIP and not ERA should be considered a more important statistic when targeting pitchers for acquisitions.
We know that BABIP fluctuates on a yearly basis, and that a pitcher has little to no control over this figure (the luck component). We also know that a strong defense can reduce the BABIP for any pitcher (the defense component). Without FIP, we would not be able to tell the difference between a player who pitches in front an above average defense and one that pitches in front of an average defense.
In addition to luck and defense, the order in which events occur has a high impact on the runs allowed by a pitcher (the sequencing component). The following example from FanGraphs perfectly explains this phenomena.
“If you had two outs and allow a single, single, home run, and out, you have just allowed three runs. If you have two outs and allow a home run, single, single, and then an out, you have just allowed one run, even though the four events were identical.”
As we can see, sequencing plays an important role in determining the ERA of a pitcher. Luck, defense, and sequencing all are capable of clouding our judgment on a pitcher’s ability. This is what Voros McCracken sought to correct, and his theory allowed Tom Tango [side note here, buy THE BOOK, it is awesome] to develop the following formula for calculating FIP.
FIP = [(13*HR)+(3*(BB+HBP)) – (2*K)]/IP + Constant
This formula assumes that the pitcher has experienced league average results on BABIP and timing. These assumptions are what allow us to strip luck, defense, and sequencing out of the equation, and allow for a level comparison of pitching performance. The constant (generally around 3.10) is used to scale FIP so that it can be read like ERA. The benchmarks for FIP are the same as the benchmarks for how we evaluate ERA. (For example, below 3.00 is excellent, 3.80 is average, etc.)
Another interesting fact about FIP is that, unlike ERA, it only accounts for events that occur while a pitcher is on the mound. Say a starting pitcher is removed from a game with 2 outs in the 6th inning, and has left 2 men on base. The reliever enters the game and gives up a home run. Under ERA calculations, two of those runs would be counted against the starting pitcher, even though he wasn’t pitching to the guy who hit the home run. However, FIP takes this scenario out of the equation by eliminating runs allowed [RA] from its inputs.
Full disclosure requires that I mention that FIP is not a perfect stat, but quite frankly, no stat is. There will be fluctuations from year to year as pitchers either over or underperform their FIP statistics. However, over the long hall, the majority of pitchers will see very little difference between their FIP and ERA statistics. There are some factors, such as having a strong ability to control the running game, or being able to generate fly balls as opposed to line drives, that make pitchers more likely to beat their FIP. However, this does not make FIP a flawed statistic. There are very few scenarios in which a pitcher outperforms his FIP consistently.
FIP is a great statistic and was considered a breakthrough in pitching analysis. It is more of a predictor of future performance than it is a tool to measure current and small sample sizes. Like all statistics, the higher the sample size, the more accurate your data can be [up to a certain point]. However, even with a season’s worth of data, FIP struggles to account for how sensitive the fly ball to home run ratio can be for a pitcher. Thus, xFIP was born.
Expected Fielding Independent Pitching - xFIP
Expected Fielding Independent Pitching (xFIP) was developed by Dave Studeman from The Hardball Times as an improvement of FIP. The only difference lies in Studeman’s finding that home run rates are unstable over time and generally fluctuate around league average, much like BABIP. As such, Strudeman took the same FIP formula, but replaced the number of home runs a pitcher gave up with the number of home runs he should have given up according to the league average HR/FB% rate.
xFIP = [(13*(Flyballs*League-Average HR/FB rate))+(3*(BB+HBP))-(2*K)]/IP + Constant
xFIP does not attempt to say that pitchers are not responsible for the home runs that they give up. It merely points out the performance the pitcher would have had if he had given up home runs at the league average HR/FB rate. This makes xFIP a better indicator of future performance than FIP because it goes one step further by turning another variable in the equation into a constant. [However, FanGraphs utilizes FIP in its WAR calculation]. xFIP has one of the highest correlations with future ERA performance, and only FanGraph’s own SIERA is higher.
Analyzing a Pitcher
While FIP and xFIP are commonly used throughout the statistical analysis community, there are some who oppose the statistics. As important as these statistics have become in evaluating pitching talent, they should not be considered the “end all, be all” absolute measures of performance. Indeed, when examining pitching performance, it may be best to used FIP and xFIP as starting points in your analysis. Once you have found pitchers who are outperforming their ERA, the next step is to ask, why? At this point, you can start to examine the pitcher at a more in-depth level and build a more accurate profile of your potential acquisition targets [this analysis structure works well in fantasy baseball too]. As with any analysis, the devil is in the details, and statistics like FIP and xFIP offer great starting points in pitcher analysis.
You can read other Back to the Basics articles here.
Jeff Meehan is a JD/MBA candidate at Suffolk University and previously interned in the Boston Red Sox legal department.