This article is the first in three exploring ways to evaluate batters for fantasy baseball using Sabermetrics. Part One will examine what makes up the commonly used statistic OPS (On-base Plus Slugging) and why it is a flawed statistical measure for fantasy baseball. Part Two will explain wOBA (weighted On-Base Average) and explore its usefulness. Part three will introduce xwOBA (Expected Weighted On-Base Average) and compare the difference with wOBA. The comparison will show hitters who have over or under-performed previously and how their projected performance might regress or progress toward the mean (a statistical term that references the average or expectation based on a data set).
Please note that I have not invented or created any of the methodologies I will present and am merely trying to coalesce information already available in an easily digestible format that pertains to fantasy baseball.
I realize the next section may be dull and unnecessary for those with more advanced baseball statistics knowledge. If you feel this is the case for you, please skip down to the “Why OPS is a Flawed Statistic” section.
Sabermetrics and History of Stats leading to OPS
For those unfamiliar with Sabermetrics, here is a brief rundown (followed by a history of how OPS came to be):
The famed baseball analyst, Bill James, coined the phrase “Sabermetrics” in 1980 to represent “the search for objective knowledge about baseball” and was named in honor of the Society of Baseball Research (SABR). Sabermetrics has been around since the origin of scorecards, and new statistical models are being formulated and refined yearly. If you don’t know anything about the subject, the book/film Moneyball is an excellent start to understand its effect on modern-day baseball.
To best understand OPS, we have first to realize how it originated and became popular. For the better part of history, Batting Average (AVG) was the defining stat for hitters. AVG is simply Hits (H) divided by At-Bats (AB). Unfortunately, it excludes Walks (Base on Balls or BB), Hit By Pitch (HBP), and Sacrifice Flys (SF). The reason for these exclusions was simple – they don’t count as ABs. AVG also treated all hits equally, be it a single or home run. Easy to see why AVG is considered inadequate, and people desired more. To better evaluate a batter’s ability to get on base, On-Base Percentage (OBP) was created to include the excluded Plate Appearances (PA) events (BB, HBP, and SF), but it still kept all hits equal. People then wanted to know how well a player hit when at the plate, and then someone created Slugging Percentage (SLG) to differentiate types of hits from one another (singles are worth one point, a double worth two, a triple was worth three, and a home run is worth four points). Please note, many advanced statistical studies have shown that these weights are not proportional to expected run production (more about this later) and thus inherently flawed. Also, SLG fails to include walks and other PAs that did not qualify as ABs.
Eventually, someone got the idea to combine/add OBP and SLG together and create On-Base Plus Slugging (OPS). This stat became a relatively mainstream statistic about 20 years ago because it was an easily digestible combination of two somewhat familiar stats and was often promoted by television broadcasts. Let’s face facts, OPS is simple to use, is everywhere, and relatively easy to understand, even for a statistical novice. If you want to know how often a batter gets on base and how successful his hits are, OPS is for you.
Why OPS is a Flawed Statistic
There are two significant problems with OPS, and they are both mathematical. The first issue is the calculation of the individual components of OPS. A perfect OBP is 1.000, and an ideal SLG is 4.000. To get an ideal OBP, a hitter needs to get on base 100% of the time; to get a perfect SLG, a hitter would have to hit a Home Run every time at-bat. Most would agree that a good OBP is between .350-.375, and the desired SLG is anything above .425. In an essay entitled “96 Families of Hitters”, Bill James calculated that an OPS of .767 or better was considered above average. It seems a useful statistic to use on the surface, right?
Remember basic arithmetic when you were taught that when adding fractions, you had to have common denominators? Well, OBP and SLG are both percentage expressions of fractions, and they have different denominators. OPS uses PAs, and SLG uses ABs. Mathematically speaking, you can’t add these two stats together. Adding OBP and SLG would require them to be calculated on the same “playing field,” as you will, and they are not – the basis for their calculation is quite different. Inconsistencies between OBP and SLG calculations are the most significant reasons statisticians worked to find a better Sabermetric measure over the years.
The second issue with OPS is also mathematical. By adding OBP and SLG together, you are treating them as equals. But the two are far from equal. Since OBP treats all hits equally and SLG does not, more weight is given to SLG in OPS. A Power hitter has a much better chance to raise his OPS than a contact hitter by only slugging extra-base hits. But how do SLG and OBP correlate to actual run production? The math here gets very complicated. I will save you the headache I suffered trying to understand how linear weights given to different types of batter’s box events change everything you thought you knew about baseball and its ability to predict run outcomes. The truth is, OPS is mathematically incompatible with linear weights (a subject to be discussed further in Part Two). In the end, I found various calculations from a multitude of statisticians that reason OBP is worth anywhere between 1.7 – 2.0 times its face value in comparison with SLG to make OPS truly equal amongst the two combined stats.
In short, OPS is fundamentally flawed. Two batters with similar OPS numbers require a more in-depth analysis to determine their fantasy worth. The contact hitter is likely to be a better hitter than the slugger (who has a similar OPS) because of the imbalanced weight inherent in the system. The real simplicity of OPS is a statistical mirage. A fantasy baseball General Manager seeking a statistical oasis will end up endlessly trekking across a desert of numbers that will mislead them in evaluating hitters when drafting their teams.
In my next article, I will introduce Weighted On-Base Percentage and explain why I (and many others) feel it is the best sabermetric measure of a hitters worth.
Connect with me on Social Media!
Twitter: @basiegel68 / Facebook: basiegel68 / Instagram: basiegel68