What is fWAR vs bWAR? Definition, Formula, and Example
fWAR and bWAR are the two public versions of Wins Above Replacement — FanGraphs' fWAR builds pitcher value from FIP, while Baseball-Reference's bWAR builds it from actual runs allowed, so the same pitcher can have very different WAR totals.
What is fWAR vs bWAR?
Wins Above Replacement (WAR) is a framework, not a single number. The two dominant public implementations are fWAR (FanGraphs) and bWAR (Baseball-Reference, sometimes called rWAR). Both try to answer "how many more wins is this player worth than a freely available replacement?" but they disagree on how to measure pitching. For position players the two systems are nearly identical; for pitchers they can differ by a win or more in a single season, which is why MVP debates cite both.
How Each Version is Calculated
For hitters, both systems combine batting (fWAR uses wOBA-based wRAA; bWAR uses its own runs-batting metric), baserunning, fielding (fWAR uses UZR; bWAR uses DRS), a positional adjustment, and a replacement-level baseline.
For pitchers, the paths diverge:
- fWAR = (FIP-based runs above average, park- and league-adjusted, scaled to innings) ÷ runs per win. It credits pitchers only for strikeouts, walks, hit-by-pitches, and home runs — outcomes independent of defense.
- bWAR = actual runs allowed (RA9), adjusted for the quality of the defense behind the pitcher, park, league, and opposition, then converted to wins.
So fWAR asks "how well did the pitcher control what he controls?" while bWAR asks "how many runs actually crossed the plate, given context?"
Worked Example: When the Two Disagree
Consider a pitcher with a 3.10 ERA but a 3.90 FIP over 200 innings — say a fly-ball pitcher who stranded an unusual number of runners. His bWAR might come in around 5.5 because bWAR rewards the actual run prevention. His fWAR might be only 3.5, because FIP says his underlying skills were closer to league average and the strand rate was partly fortune. This pattern shows up constantly: in 2021, Robbie Ray's Cy Young season rated higher in fWAR (elite strikeout rate, great FIP) than bWAR, while pitchers with soft-contact profiles like Kyle Hendricks historically grade better in bWAR than fWAR. A rule of thumb: a gap larger than 1.0 win between the two flags either unusual sequencing luck or a pitcher whose skill set one model captures and the other misses.
Why fWAR vs bWAR Matters
Front offices treat the gap as diagnostic. A pitcher whose bWAR consistently exceeds his fWAR (Hendricks, early-career Dallas Keuchel) may have real contact-management skill that FIP-based models underrate. For Hall of Fame and MVP debates, both are cited — Baseball-Reference's WAR is the default on player pages, while FanGraphs' version dominates analytical writing. In fantasy baseball, fWAR's components (K, BB, HR) are more stable year to year, making fWAR the better forecasting input.
In Legends Deck: pitcher card ratings blend both philosophies — stuff-driven metrics (strikeout and walk rates) set the card's baseline, while actual run prevention from the rated season shapes in-game simulation outcomes, so a card reflects both dominance and results.
Limitations and Common Misconceptions
Neither is "correct" — they answer different questions. The common mistake is treating a 0.8-win gap as a meaningful disagreement about a player's true talent; WAR's error bars are larger than that. Another misconception: the two systems are not interchangeable when summing career value, and mixing them in one argument ("his 60 bWAR and 55 fWAR average to 57.5") is not standard practice. Also note a third version, Baseball Prospectus's WARP, uses DRA and produces yet another set of numbers.