What is Pitching+? Definition, Formula, and Example
Pitching+ is a FanGraphs model that combines Stuff+ (pitch shape quality) and Location+ (command quality) into a single 100-scaled metric, where 100 is league average and every point above represents roughly 1% better expected run prevention.
Plain-English definition
Pitching+ is a sabermetric model developed by Eno Sarris, Max Bay, and the FanGraphs team that grades a pitcher's overall expected effectiveness by combining two component models — Stuff+ (how nasty the pitch itself is) and Location+ (how well it was located given the count and batter handedness). It is scaled around 100 = MLB average, with every point above or below representing approximately 1% better or worse than league average run prevention. A 110 Pitching+ means the pitcher's underlying pitch quality and command should produce results 10% better than average; an 85 means 15% worse.
How Pitching+ is calculated
The model works in three layers:
1. Stuff+ — a gradient-boosted regression model trained on Statcast inputs: release velocity, spin rate, induced vertical break, horizontal break, extension, release point, and the velocity/movement differential between the pitch and the pitcher's primary fastball. The output is each pitch's predicted run value, then scaled.
2. Location+ — a separate model that grades pitch location given count, batter handedness, and pitch type. A fastball at the top of the zone in 0-2 grades very differently than the same pitch in 3-1.
3. Pitching+ — a weighted blend of Stuff+ and Location+, with location weighted more heavily for command artists and stuff weighted more heavily for power arms. The exact weighting reflects the empirical contribution of each component to run prevention in the training set.
The formula is:
Pitching+ ≈ (Stuff+ × w_s) + (Location+ × w_l), with weights tuned per pitch type so the final number predicts ERA over a large sample better than either component alone.
Worked example
Tarik Skubal's 2024 Cy Young season featured a Pitching+ of approximately 111 — driven by an elite 115 Stuff+ (top-5 in MLB on his fastball-changeup combination) paired with a slightly above-average 102 Location+. Compare to Logan Webb, whose 2024 line showed a 92 Stuff+ but a 107 Location+, producing a Pitching+ around 100 — exactly average expected performance, which mapped neatly onto his 3.47 ERA. The two pitchers achieved similar results through opposite means: Skubal misses bats with shape, Webb misses barrels with command and sequencing.
At the bottom of the league, replacement-level swingmen typically sit at 88-93 Pitching+, which is why they cycle between Triple-A and the back of bullpens.
Why it matters
Pitching+ is one of the strongest predictive ERA estimators in public baseball analytics, often outperforming FIP and xERA in next-season projection because it operates at the per-pitch level rather than aggregating outcomes. Front offices use Stuff+, Location+, and Pitching+ as standard inputs in trade evaluation, free-agent contract modeling, and amateur scouting. Player-development staff use the components diagnostically: a pitcher with high Stuff+ but low Location+ has a clear development path (work on command), while a high-Location+, low-Stuff+ profile suggests pitch-design intervention.
Limitations and common misconceptions
Pitching+ is model-dependent: different vendors (FanGraphs, Driveline's PitchingBot, the proprietary models used by clubs) train on different inputs and produce slightly different numbers for the same pitcher. The model does not capture sequencing (what pitch followed what), pitch tunneling between consecutive offerings, or batter-specific deception — pitchers who excel at unpredictability can outperform their Pitching+. The model also has reduced reliability in small samples: under ~500 pitches, regression toward the mean is heavy. And because Stuff+ rewards velocity and movement, deception specialists like Kyle Hendricks have historically been undervalued by raw Stuff+ even when their Location+ partially compensates.
Related terms
In Legends Deck
Legends Deck pitcher cards carry separate Stuff and Command ratings derived from the underlying Statcast inputs that feed Stuff+ and Location+, plus a composite Overall Pitching rating analogous to Pitching+. This separation lets you build deck strategies around either profile — a power-arm rotation that wins on whiffs, or a command-and-control staff that wins on weak contact and efficient innings.