We fans want a number to measure and compare the quality of our baseball players. Trouble is, so many of the numbers we’re comfortable with don’t really give us a full picture.

Batting average (abbreviated as BA or AVG), though traditional and ubiquitous, is perhaps the most flawed of all. On-base percentage (OBP) is more useful, as is slugging percentage (SLG), though the attempt to give us a holistic look by combining those two into one measure, OPS (literally “on-base percentage *plus *slugging”), doesn’t really work for a number of good mathematical reasons.

We rely on slash lines of AVG, OBP, SLG, despite their one dimensionality, just as we take our temperature to see if we are running a fever, knowing full well that the absence of fever does not necessarily mean we’re healthy.

RSNStats Recommends…

To learn more about baseball stats RSNStats recommends these outstanding resources:

*The Book: Playing the Percentages in Baseball* by Tom Tango, Mitchel Lichtman and Andrew Dolphin

*Baseball Between the Numbers* by Jonah Keri

*A Mathematician at the Ballpark* by Ken Ross

## wOBA: A Better Performance Metric

For evaluating both batters and pitchers, Keith Law, in his excellent book *Smart **Baseball*, recommends weighted on-base average (wOBA), a sabermetric based on linear weights that was introduced in the 2006 book *The Book: Playing the Percentages in Baseball* by Tom Tango, Mitchel Lichtman and Andrew Dolphin.

wOBA covers all the typical batter’s events (hits, extra base hits, walks, times hit by a pitch, and outs) and applies weights to each of those components that correspond to their impact on the game.

wOBA weights aren’t applied as simplistically as they are with SLG, which wrongly assumes that triples are worth three times more than singles, for example. Instead, weighting for wOBA is systematically determined based on the run scoring environment and updated every year. As baseball changes due to any number of external events, the weighting adjustments used in wOBA make these individual feats more valuable and, naturally, the converse is true. When, for example home runs, are more plentiful, the weighting reduces their value in the wOBA calculation.

Weighting values used in determining wOBA can be different depending on who is doing the calculating. RSNStats relies upon the weights determined by the Fangraphs web site, which makes them available for all years going back to 1871.

In 2020, the formula for wOBA was *(0.699×Unintentional BB + 0.728×HBP + 0.883×singles + 1.238×doubles + 1.558×triples + 1.979×HR) / (AB + BB – Intentional BB + SF + HBP).*

**What’s a good wOBA?** The formula for wOBA scales it to league average for on-base percentage. This is handy because if you already know what a good OBP is, you also know a good wOBA.

In 2019, the Major League average for on-base percentage is .323, so as you approach .400 you’re getting into excellent performance for both measures.

For pitchers, of course, the opposite is true: the lower the opposing wOBA, the more effective the pitcher has been.

## Understanding Expected wOBA

As with other advanced stats, there are so-called “expected” versions. You’ll recognize these because they start with the letter “x.”

In the case of wOBA, the expected version, xwOBA, can give a clearer view of a batter’s skill because, as former MLB data scientist Sam Sharpe writes, xwOBA focuses on certain “tracked skills to reach a conclusion about what would have happened to balls in play under completely average MLB game conditions.” By eliminating the effects of defense on an offensive effort, expected stats like xwOBA rely on determining a batter’s quality of contact using factors such as exit velocity and launch angle of the batted ball and the batter’s sprint speed to determine the expected outcome as opposed to the actual outcome.

You may be tempted to say you only care about a player’s *actual* performance and not what typically occurs. But since batters can’t control what happens when the ball leaves the bat, it’s useful to evaluate the player’s contribution based on what *typically* happens given a certain combination of offensive characteristics.

For example, a batter who reaches safely on a misplayed ball gets lucky. wOBA credits the player with the result. But xwOBA likely concludes that the batter’s performance, as measured by a variety of characteristics, would under average conditions, likely been an out. xwOBA, therefore, may give you a more realistic understanding of the player’s actual skills.

## More Ways to Evaluate Pitchers

Another useful metric for evaluating pitchers is FIP: Fielding Independent Pitching. This is because once a ball is put into play, there’s not much difference in talent among pitchers. As Tom Tango notes, what you see might tell you there’s a difference but much of that is random variation.

Pitching wins are really immaterial. More reliable measures for evaluating pitchers are their strikeouts, walks, hit-by-pitch and home runs allowed. These are the actions that a pitcher truly controls.

FIP abstracts away the fielding plays to give you a measure, much like Earned Run Average (ERA), of the runs allowed by what the pitcher controls most.

**What’s a good FIP?** Like ERA, the lower the FIP, the better the pitcher has been.

In 2019, the Major League average for Field Independent Pitching was 4.50.

## Learn More

There’s little reason to create a new glossary for stats here at RSNStats.com. Rather, if you’re interested to know more, check out the recommended books (above) or the explanations available on the Fangraphs, Hardball Times, and Major League Baseball web sites.

Eli Desjardins says

Awesome article. Thank you for being such an awesome twitter follow and I think your page will blow up in the next few years.

RSNStats says

Appreciate the kind words. Thanks for following along.

Michael says

You mention that FanGraphs has the weight factors going back to the 1800s. Okay. But what if I want to calculate the factors for non-MLB leagues? What is the formula for calculating the weight factors? That’s what I’m having a hard time finding.

Billy says

these measures should be statistically validated. OPS actually *performs* better than wOBP in explaining/predicting run production–so theory for wOBP is nice but ultimately not of consequence

RSNStats says

OPS is familiar, which makes it useful, and if you like it, use it. But OPS is not mathematically sound. Specifically, OBP and SLG, the two components of OPS use different denominators (OBP uses plate appearances, SLG uses at-bats). More importantly, though, adding OBP and SLG together just doesn’t make sense. It overweighs the value of SLG, since SLG values are often much higher than OBP values.