What is FIP in Baseball?

FIP in baseball is the result of a relatively new analytics push to separate pitcher impacts from defense impacts. FIP stands for Fielding Independent Pitching. Sometimes good defense and fielding can bail out bad pitching. Sometimes the opposite happens and the defense hurts the pitcher.

The point of FIP in baseball is to really get to the heart of the matter and take defense completely out of the equation. The result is a stat that measures pitcher impacts and only pitcher impacts.

In this article we’ll take a deep dive on FIP in baseball. We’ll also look more broadly at the field of DIPS (defense independent pitching statistics) which all try to get at the same idea. After this we’ll compare FIP to other stats that try to do the same thing. Finally, we’ll talk at a higher level about how useful the stat is to evaluate baseball talent.

To receive email updates when new articles are posted, use the subscription form below!

What is FIP in Baseball?

One of my favorite topics in sports analytics is the credit assignment problem. Depending on the sport, there are at least 10 players on the field at a given time. Who is responsible for the outcome of the play?

We know that one player can’t do everything, but we also know that not everyone is equally important. If a pitcher gives up a single but an error turns it into a double, is the pitcher responsible for 50% of that negative outcome? More? Less? I’ve always claimed this is the hardest and most important problem in sports analytics. FIP in baseball is a way to tackle it.

FIP in baseball isolates a pitcher’s impacts by looking at the plays that they were 100% responsible for. Because of this, the only stats that FIP incorporates are strikeouts, walks, HBP, and home runs. These four events have no fielding involved at all!

If a strikeout happens, it is because the pitcher made it happen. If a home run happens, it wasn’t because of any good or bad fielding; the defense never touched the ball. The same is true on walks and players hit by a pitch. These things fall squarely on a pitcher’s shoulders.

Measuring these four stats and combining them in a meaningful way is the basis for FIP in baseball. The next section contains the fielding independent pitching formula.

What is a Good FIP in Baseball?

Before even understanding the formula for a stat, understanding what different numbers mean is key. Everyone knows how to interpret ERA, for example, but far fewer folks know how to interpret what FIP numbers mean.

It turns out, though, that the numbers are about the same. This isn’t a coincidence, it is designed to be this way. An ERA of 3 is pretty good, a FIP of 3 is pretty good. An ERA of 6 is bad, a FIP of 6 is awful. The graphic below shows how we break down the ranges of FIP values.

More information about FIP values can be found on baseball reference.

The FIP Formula

In the FIP formula below, the following abbreviations are used:

• BB: walks allowed
• HBP: Batters hit by pitch
• HR: Home runs allowed
• K: Strikeouts
• IP: Innings Pitched
• c_{FIP}: “The FIP Constant” (will explain later)

With this notation, the FIP formula is FIP=\frac{13HR+3(BB+HBP)-2K}{IP}+c_{FIP} . Let’s dig into this formula and point out some features.

First, what is going on with the FIP constant c_{FIP} ? In general, it is desirable for a stat to be “interpretable”. For example, earned run average is easy to understand; it tells you how many runs a pitcher allows over the course of a normal game.

Just combining home runs, walks, HBP, and strikeouts doesn’t tell us much. The point of adding the FIP constant is to make FIP look a lot like ERA. If a pitcher has a 4 FIP, then their ERA is expected to be 4. If a pitcher has 8 FIP, then ERA should be around 8.

Also notice that FIP is a rate state, it measures “things allowed” per inning pitched. This means that simply pitching more or less innings shouldn’t change a pitcher’s fielding independent pitching.

The last thing we notice is the difference between the multipliers. The constants roughly represent the relative “goodness” or “badness of an event”. Remember from above that larger FIP is bad. In the FIP formula, HR shave a 13 multiplier while walks have a 3. This represents the obvious fact that homeruns are worse than walks. It also tells us that home runs are about 4 times as bad as walks.

In the next few sections we’ll talk about fielding independent pitching from a higher level and think about how it may be useful.

Is FIP a Good Stat?

The first question we should ask is “Is FIP a Good Stat”? The short answer is…kind of? It has value and accurately identifies good and bad pitchers. But I don’t think it does it as well as some other stats.

Fielding independent pitching is based on a very good idea. It has a noble purpose in theory. FIP in baseball is predicated on the hypothesis that “to accurately measure how good a player is, you need to isolate their impact on the game from everyone else’s impact on the game”. I agree whole heartedly.

It is the methods of doing this that I don’t particularly love. The next logical leap that fielding independent pitching makes is that “the only way to divorce a pitcher’s impacts from the rest of the fielders is to only look at plays where the fielders didn’t participate”. This is why FIP only uses walks, home runs, strikeouts, and HBP.

I think it is entirely possible to somehow measure that a routine popup is a good outcome. I think it is possible (and in fact very easy) to measure that a clear single turned into a double by an error should only count as a single against the pitcher’s record.

Baseball is lucky because there is not nearly as much gray area as other sports. It is straightforward to isolate which impact which players were responsible for. And I don’t think that FIP does this.

So is FIP a good stat? I’m not totally sold on it. I think it can contribute to an overall picture of a pitcher. Kind of like how BABIP can contribute to an overall picture of a hitter’s quality but is not valuable on its own.

In the next section, I’ll include some more of my gripes with FIP.

Problems with Fielding Independent Pitching

I spy a few other problems with fielding independent pitching as currently defined.

Problem 1: Doesn’t Normalize for Hitters

To me, this is the biggest problem with FIP in baseball and is, frankly, unforgivable. The entire point of fielding independent pitching is to isolate the actions of the pitcher so we can understand how good they are. This stat works by only looking at events where fielders aren’t involved. The idea is that this way only the pitcher is involved in the play.

Except that isn’t the case. No matter what, the hitter is always involved in the play. We can’t isolate ONLY the pitcher’s contribution because the hitter always has a say in the outcome of an at bat. Sometimes hitters can just force a homerun on sheer force of will. Sometimes they strikeout because they haven’t been seeing the ball.

On the other hand, what if you happen to be a starter in the hardest division in baseball so that lots of your innings are against great offenses? Your FIP will naturally be larger because you faced tougher competition.

Failing to normalize fielding independent pitching for the quality of the hitter is a serious shortcoming and frankly invalidates the whole philosophy of the stat. Luckily, this is a fairly trivial fix that could be made. Think something like FIP+.

Problem 2: Certain Pitching Styles are Punished

The other major complaint I have with fielding independent pitching is that it doesn’t tell the whole story. Home runs, walks, Ks, and HBPs are not the entirety of a pitcher’s identity.

Some pitchers – often closers – are strikeout masters. They live to strike a guy out. Others don’t power past a guy, they are simply ground ball pitchers and thrive on their defense behind them. Even if these two pitchers allow the exact same number of runs, FIP will tell a different story.

FIP will boost how good a strikeout pitcher looks. It will make ground ball pitchers or guys that get lots of popouts look worse. Good stats don’t punish people for having different styles. Good stats capture the outcome without focusing on the “how”.

Problem 3: It Isn’t Maximally Accurate

My last complaint is smaller, but it deserves to be said. I want to focus on the coefficients in the equation. The fact that all the multipliers are integers (13, 3, and 2) means that they are approximations for the relative value of the various outcomes.

If these were the right numbers, there is no way they would be so well rounded. Look back at my last article about Pythagorean wins and Pythagorean expectation. The original formula used exponents of 2: nice, well rounded numbers. But it turns out when you do the math, the optimal exponent is 1.83.

So my point is that FIP is not as accurate as it could be. There is always a tradeoff between interpretability and accuracy. I claim FIP already sacrifices interpretability, especially because it uses the weird FIP constant. If we already sacrifice an interpretable formula for a stat, why not go all the way and use the right coefficients?

Other DIPS Stats

Fielding independent pitching is only one example of a larger set of baseball stats known as “defense independent pitching stats” or DIPS. DIPS refers to the idea that taking defense out of the equation when evaluating a pitcher and how good they are is a good idea.

There are many DIPS stats including things like “DICE” and “SIERA”. I actually claim that something as simple as ERA is in the spirit of DIPS. ERA counts runs a pitcher gives up, but it subtracts the “unearned runs”. ERA discounts those runs which are clearly the defense’s fault. ERA is an example of DIPS.

Overall, DIPS stats are interesting and should be studied and expanded upon. I just don’t think that FIP is the best example of this stat. So, as popular as fielding independent pitching is, I don’t quite think it is warranted.