FIP, xFIP and SIERA: Estimating a Pitcher's Skill from What He Controls

By 4Baseball Staff · November 5, 2026 · 10 min read

What FIP, xFIP and SIERA measure, how to calculate FIP by hand with a worked example, why defense-independent pitching began with Voros McCracken, and when each metric helps or misleads.

Earned run average is one of the oldest ways of saying how well a pitcher did, and it has a real flaw: it counts everything that happened while he was on the mound, including things he did not control. A line drive that finds a fielder's glove is an out, and the same line drive hit three feet to the left is a double. FIP, short for Fielding Independent Pitching, throws out the balls that were put in play and keeps only the outcomes a pitcher decides almost on his own: strikeouts, walks, hit batters and home runs. The result is a number on the same scale as ERA, so a 3.40 FIP reads like a 3.40 ERA.

The family has two close relatives. xFIP replaces a pitcher's actual home runs with the number you would expect from his fly balls, and SIERA goes further, building a more detailed model of what produces runs. This article explains where the idea came from, how to work out FIP on a napkin, what xFIP and SIERA add, what each one misses and how to use them without being fooled. The historical record carries the counting numbers FIP needs, which is why some of the figures below are computed from it.

The idea that started it: Voros McCracken's discovery

In 2001 a researcher named Voros McCracken published an analysis that startled many people who studied baseball. He examined what happens to balls in play, the ones that are not strikeouts, walks or home runs, and found that individual pitchers differed far less in their rate of hits on those balls than anyone had assumed. A pitcher who gave up a .260 batting average on balls in play one year might give up .310 the next, with no change in his stuff. The rates for strikeouts, walks and home runs, by contrast, were stable from year to year and clearly belonged to the pitcher.

The approach became known as defense-independent pitching statistics, or DIPS. The conclusion was strong, and later research softened it. Pitchers do have some influence on balls in play, since a hitter makes weaker contact against some pitchers than others, and ground-ball and fly-ball tendencies matter. But the core claim has stood up: most of the year-to-year swing in a pitcher's hits allowed on balls in play comes from luck and from the defense behind him. The practical result is that a pitcher's earned run average has a lot of noise in it.

Tom Tango, who worked in the same sabermetric community, turned the idea into the formula known as FIP. The model is deliberately simple. It asks what a pitcher's ERA would have been if every ball in play had gone for a league-average result.

The FIP formula, step by step

The standard formula is:

FIP = (13 × HR + 3 × (BB + HBP) − 2 × K) ÷ IP + constant

HR is home runs allowed, BB is walks, HBP is hit batters, K is strikeouts and IP is innings pitched, with partial innings as thirds. Some versions leave intentional walks out of the walk total, and the constant is chosen each season so that the league's overall FIP matches the league's overall ERA.

The weights come from the run values of each event. A home run is the most damaging thing a pitcher can allow, worth far more than a walk, and a strikeout is better than an ordinary out because no runner can advance on it. The ratio 13 to 3 to 2 reflects those relative values. A home run is weighted more than four times as heavily as a walk, and each strikeout subtracts two points from the numerator.

Here is a worked example with three hypothetical pitchers and an illustrative constant of 3.10.

Pitcher A strikes out many batters and gives up few home runs, so his strikeout total pulls the numerator below zero and his FIP is lower than the constant. Pitcher B gives up 30 home runs and strikes out only 140, so the numerator is large. Pitcher C allows few home runs but has few strikeouts, and ends up in the middle. All three are invented, and the constant is only illustrative.

Why the constant exists, and how it moves

Without the constant, the formula would produce numbers on a different scale from ERA, often near zero or even negative. The constant shifts the whole scale so that the league's FIP and ERA agree in the season in question. Because ERA depends on the league's run environment, the constant changes from year to year.

The table below computes the constant from the historical team totals for selected seasons, as the league ERA minus the league's FIP-style rate before the constant, counting hit batters with walks. The historical walk totals include intentional walks, so published constants will be slightly different, and the figures are for illustration.

Two lessons come from the table. First, the constant has risen from about 2.4 in the Year of the Pitcher to about 3.1 or 3.2 in recent seasons, so a number such as 3.10 is a good rough value for modern baseball but not for the 1960s. Second, the strikeout column shows how much the game has changed. League strikeouts per nine innings rose from 5.89 in 1968 to 8.49 in 2025. A pitcher with a certain strikeout rate in one era is not the same pitcher in another.

What FIP is good at, and what it misses

FIP is a better predictor of a pitcher's next-season ERA than his ERA is, in most published comparisons, because it strips out the noisy portion. It is most useful in three situations.

1. Spotting luck. A pitcher whose ERA is well below his FIP may have been helped by a good defense, a low rate of hits on balls in play or a high percentage of runners stranded. His ERA is a candidate to rise. A pitcher whose ERA is well above his FIP may have been unlucky, and his ERA is a candidate to fall.
2. Comparing pitchers on different teams. Because FIP removes defense, a pitcher on a team with a weak infield is not penalized for it, and a pitcher on a team of great gloves is not rewarded for it.
3. Short samples. Strikeout and walk rates stabilize faster than hit rates on balls in play. Over a few weeks, a pitcher's FIP is more informative than his ERA.

It also has real limits. FIP assumes that every pitcher has the same control over balls in play, and that is not exactly true. Knuckleball pitchers, pitchers who consistently induce weak contact or pop-ups, and pitchers with strong ground-ball tendencies can beat their FIP year after year. FIP also ignores how runs are sequenced: a pitcher who gets a strikeout with the bases loaded and one who gets one with the bases empty are credited equally, even though the value to the team is different, which is a point that win probability and leverage handle better.

Finally, FIP does not know where the games were played. A home run total in a park that gives up many home runs is scored the same as the same total in a park that suppresses them, which is a significant distortion for pitchers at the extremes.

xFIP: replacing the home runs

The home run is the most volatile of FIP's inputs. Whether a fly ball leaves the park depends on the wind, the ballpark, the temperature and the batter's swing, and in most pitchers' careers the share of fly balls that become home runs (HR/FB) swings from year to year more than their fly-ball rate does. xFIP, for expected FIP, is built to deal with that.

The idea is to take the pitcher's number of fly balls and multiply it by the league-average HR/FB rate, which gives an estimate of the home runs he should have allowed. The formula then runs the same way as FIP:

xFIP = (13 × (FB × league HR/FB rate) + 3 × (BB + HBP) − 2 × K) ÷ IP + constant

The constant is again adjusted so that xFIP matches league ERA. In practice xFIP has an even smaller random component than FIP, and it is often a better predictor of future ERA for pitchers who have had an unusually high or low home run rate.

xFIP pays a price for its assumption. If a pitcher plays in a park that suppresses home runs, or if he really does have a skill at limiting the damage on fly balls, xFIP will understate how good he is. Pitchers who work at Oracle Park or Coors Field, for example, have home run rates that depend on the park, and xFIP treats them as if the park were average. The historical record does not carry fly-ball counts, so xFIP cannot be computed from it. It needs batted-ball data from other sources.

SIERA: a more detailed model

SIERA stands for Skill-Interactive Earned Run Average. It was developed by Matt Swartz and Eno Sarris and published in 2010. Where FIP and xFIP are fixed formulas with simple weights, SIERA was built by fitting a statistical model to a large number of pitching seasons, with the goal of predicting a pitcher's ERA.

Two features set it apart. First, it includes the pitcher's ground-ball and fly-ball tendencies, as well as strikeouts and walks, so a ground-ball pitcher is credited for the weak contact he tends to produce. Second, it allows the inputs to interact, which is where the name comes from. The effect of one skill can depend on the level of another: a walk, for example, tends to be less costly for a pitcher who rarely lets hitters put the ball in play than for one who allows a lot of contact. A fixed linear formula cannot represent effects like that, and SIERA includes squared and cross terms so that it can.

The result is a formula that most people do not calculate by hand. It lives on stat sites, and its output is on the ERA scale. Its authors reported that it forecasts future ERA slightly better than xFIP, although the margin is small and depends on the sample.

Which one to use

The three measures answer slightly different questions, and the right one depends on what you want to know.

Beyond that, a few rules of thumb help.

• Use ERA to describe what a pitcher did and FIP to describe what he did that he controlled.
• Use xFIP or SIERA when you want to know what is likely to happen next.
• When the three agree, believe the answer. When they disagree, look for the reason: a high BABIP, a low strand rate, a park effect or an unusual home run rate.
• Use more than one season. A single year of any of these metrics is a noisy estimate of skill.

Wins Above Replacement for pitchers is built on one of these ideas in some versions. One widely used version of WAR is based on FIP, while another is based on runs allowed adjusted for defense. The difference is explained in the guide to WAR.

Common mistakes

Treating FIP as a measure of quality instead of a measure of process. A pitcher with a poor ERA and a strong FIP is not automatically a good pitcher who has been unlucky. He may allow hard contact that FIP does not see.

Comparing across eras without context. Because the constant and the strikeout rate change, a 3.50 FIP means different things in 1968 and 2025. The cleaner way is to compare a pitcher to his league, as ERA and ERA+ do for earned run average.

Using small samples. A month of FIP is mostly noise. Strikeouts and walks become informative first, which is the idea behind looking at strikeout and walk rates early in a season.

Assuming a gap between ERA and FIP is always luck. Some pitchers really do manage contact better than average, and FIP undervalues them. Check the pitcher's history over several seasons before deciding that a gap is luck.

Ignoring WHIP. WHIP measures baserunners per inning and includes hits on balls in play. It is a useful companion because it shows how much traffic a pitcher allows, whatever the cause.

Common questions

Is FIP better than ERA? It is better for prediction and for comparing pitchers on different defenses, but ERA is the number that determines runs, so it is the right measure of what happened.

Why 13, 3 and 2? They are rounded weights that reflect the run value of each event: a home run is the worst outcome for a pitcher, a walk or hit batter costs a smaller amount and a strikeout is the only one of the three that helps.

Can I calculate FIP from the historical record? Yes. Home runs, walks, hit batters, strikeouts and innings are all in the pitching tables, though you need the season's constant from a source that publishes it, or you can compute it yourself from league totals as in the table above.

What is a good FIP? It is a number on the ERA scale, so compare it with the league's ERA in that season. Better than the league average is good, and about a run better is excellent.

Do FIP and SIERA replace scouting? No. They describe results from the numbers; they do not tell you why a pitch works, or whether a young pitcher is changing.

Where to go next

Read WHIP explained for the other great pitching rate statistic, and WAR explained for how these ideas feed into overall value. To see the numbers behind it, the career leaders pages list strikeouts and walks, and the stats and analytics topic page gathers more explainers.