While the baseball community has largely advanced past the utilization of purely descriptive traditional statistics to evaluate player performance, occasionally we witness the resurrection of their futility. Pitching win-loss records have largely been debunked as a poor tool for evaluating a player’s performance, for after all, pitchers cannot control the runs their teammates produce. The Orioles’ Chris Tillman, however, earned an All-Star nod undoubtedly due to his sterling 11-3 clouding a pedestrian earned run average of 3.95 and even worse figures 4.95 FIP (fielding independent pitching), 4.23 xFIP (expected FIP), and 0.6 WAR (wins above replacement) that are not only not All-Star worthy, but also simply below average. Progress remains slow.
Baseball executives and other individuals who harness the ability to impact the product on the field, however, recognize the utility of such sabermetrics, with the Oakland Athletics a perpetual example of its success. While traditionalists – many of whom reside in the mainstream media – continue to fight against the revolution, to be a fully informed fan of the game it is important to recognize its complexities and be willing to adapt to new developments in its understanding. Although we have come a long way, there still remains many ways to improve, but by starting to evaluate players by some of the methods described in this article, you will have a greater understanding than the average fan that relies on traditional statistics.
So what, then, are sabermetrics? As pioneer Bill James defines, sabermetrics involve the “search for objective knowledge about baseball.” Sounds simple, right? We’ll narrow it down to describe sabermetrics as the utilization of statistical tools that seek to evaluate and objectively predict the true nature of a players’ performance.
While traditional statistics are descriptive, most sabermetrics are predictive. Sabermetrics seek to compensate for the unpredictability of luck. Over time, a certain amount of ground balls tend to turn to outs, a portion of fly balls tend to result in home runs, etc. Some players may consistently over time reach base on more grounders due to speed or other factors, but much of the time such results cannot be reliably predicted and tend to regress to the mean over time. Some players get lucky or unlucky in comparison to those trends due to factors such as the quality of the defenders behind them or park size. Sabermetrics thus attempt to correct for statistical anomalies for those who experience an unlucky amount of ground balls squeak through the infield, fly balls that leave the yard, or the range of the pitchers’ defenders, for example. Likewise, players who perform abnormally compared to their minor league track record, career in the majors, and in comparison to league average in a variety of statistics, are more or less due to regress to the mean.
The idea is that a player’s luck, or lack thereof, is most likely to be average moving forward. Is a pitcher with a limited-ranged Derek Jeter at short with the same strikeout rate, ground ball rate, etc. a worse pitcher than another with an exceptional defender like Andrelton Simmons at short getting to more ground balls and creating more outs? If you couldn’t tell where I was going with that hypothetical, the idea is that the pitchers should be about the same, even if the former’s ERA exceeds the latter’s due to more runners advancing with a worse fielder. Sabermetrics attempt to correct for these differences, and teams may be able to utilize such statistics to target players who perform poorly based on traditional numbers but whose skills suggest improvement. For pitchers with high earned run averages yet low fielding independent pitching scores (more on “FIP” later), adding better quality defenders may instigate an improvement in earned run average, a purely descriptive statistic, in the future.
Likewise, executives utilize sabermetrics to find the factors that contribute to the ultimate goal – wins. While simple logic leads us to believe that a higher batting average leads to more runs with more base runners, and therefore more wins, other statistics are more accurate predictors of wins. Michael Lewis’ Moneyball highlighted Oakland Athletics General Manager Billy Beane’s ability to recognize a market inefficiency – that on-base percentage correlated more closely with wins than batting average – in order to produce a winner with players paid at a market value less than their actual worth toward wins, based on mathematical regressions by Paul DePodesta and others. Since the collective baseball world did not realize that, Beane was able to find players undervalued on the market. Beane still appears to be ahead of the rest of the league with his low budget squad leading the AL West after likewise emerging victorious in 2012, ahead of the superstar laden Rangers and Angels rosters by finding the next market inefficiency, specifically commandeering a deep bench with depth throughout the roster in place of superstars, a similar approach taken in the Boston Red Sox’s incredible turnaround this season.
So throughout the rest of the article, we will catalog many of the quick and easy advanced statistics to utilize to make you a smarter fan of the national pastime.
Offense: Weighted On-Base Average (wOBA) – An Alternative to On-Base Plus Slugging Percentage (OPS)
OPS, on base percentage plus slugging percentage, has developed into a popular barometer for offensive success, seemingly replacing good ole batting average as the most preferred offensive average. While OPS is a more predictive statistic than batting average since it incorporates additional offensive factors (batting average ignores walks and does not delineate between a single versus a home run, for instance), it still contains crucial flaws. Not only does it mathematically lack consistency given that it involves simple addition between two unlike numbers with different denominators (on-base percentage incorporates plate appearances, while slugging percentage uses at-bats for their averages), but it assumes that the two numbers are equally valuable to OPS. Two batters each with OPS averages of .750, for instance, may not be the same – one could have a .300 on-base percentage with a .450 slugging percentage, while the second could have a .350 on-base percentage and .400 slugging percentage. According to research by Tom Tango and subsequent analysis by others, each “point” of on-base percentage is actually approximately 1.8 times more valuable than the slugging percentage component of OPS. For further analysis on how Tango arrived at his conclusions, click on this link here to Tango’s article from 2007 on the website “THE BOOK.
Thus, a modified OPS would seem to be a better indicator for success than the OPS we are accustomed to. Fortunately, such a metric has been created, though it more closely resembles numbers seen from on-base percentage, primarily to scale to a number most people can easily evaluate and compare players (OPS is still relatively new to the mainstream). For instance, we perceive an OBP value of .350 to be above average, and thus a weighted on-base average (wOBA) of .350 is likewise to be considered the same.
But what is wOBA and how is it calculated? I will defer to Fangraph’s Dave Cameron to provide an eloquent explanation: “wOBA is based on a simple concept: Not all hits are created equal. … Weighted On-Base Average combines all the different aspects of hitting into one metric, weighting each of them in proportion to their actual run value. While batting average, on-base percentage, and slugging percentage fall short in accuracy and scope, wOBA measures and captures offensive value more accurately and comprehensively.” Likewise, the league average wOBA is set to closely resemble the league average OBP, and thus fluctuates every season. As a general rule of thumb (adopted from Fangraphs), a wOBA of .320 is considered average, .340 above average, .370 great, and .400 excellent, while values like .310 are below average, .300 poor, and .290 poor.
Interested in the formula? While it fluctuates every year, the formula for the 2012 season was the following:
wOBA = (0.691×uBB + 0.722×HBP + 0.884×1B + 1.257×2B + 1.593×3B +
2.058×HR) / (AB + BB – IBB + SF + HBP)
While you do not really need to know the complexities of determining the formula, the idea is simple – wOBA is an excellent all-encompassing offensive statistic to evaluate players’ contributions from the plate (note the usage of “from the plate” – wOBA does NOT incorporate base stealing and baserunning metrics into the value. We will get to that later). Given their strong on-base and slugging skills, it is no surprise that Miguel Cabrera, Chris Davis, and Mike Trout lead the league with wOBA values of .478, .443, and .428, respectively. Ultimately, by utilizing weighted on-base average to discuss players’ purely offensive contributions, you will be smarter than peers that still use batting average, on-base percentage, or even OPS.
Weighted Runs Above Average (wRAA) and Weighted Runs Created (wRC)
Another vital advanced statistic for calculating offensive output involves two similar statistics, weighted runs above average (wRAA) and weighted runs created (wRC) (their difference is largely in the scaling of the numbers). Essentially, both statistics demonstrate the amount of runs created by a player compared to league average.
Given that we can directly use wOBA to calculate wRAA (and that wRAA is utilized to calculate the famous all encompassing WAR – more on that later), we can find a player’s value with the following formula:
wRAA = ((wOBA – league wOBA) / wOBA scale) × Plate Appearances
(The wOBA scale is a constant that varies year to year)
Basically, values are self-explanatory; a wRAA of 40 means a player creates 40 runs above league average. Using Fangraphs’ scale, 40 is considered excellent, while a value such as -20 is considered awful. While we will return to it later, each run above average contributes +0.1 to their wins above replacement (WAR) value, working off the idea that each game generally on average features ten runs scored.
Other Offensive Notes
There are countless other statistics that will allow you to escape ignorance when discussing a player’s offensive performance, but there are a few other worthy factors to briefly touch upon. Isolated power (ISO) is a short hand statistic that separates the effects of batting average on a player’s slugging percentage as simply the difference between their slugging percentage and average. ISO allows us to easily recognize the impact or frequency with which a hitter reaches extra bases when they do generate hits. A player with a lower batting average than another with the same slugging percentage may be perceived as having more power, at least in the samples provided.
Strikeout and walk rates are also excellent metrics to view to project player performance, to a certain extent. A player with a high batting average yet a high strikeout rate may be prone to a regression in their average since they put fewer balls in play. Such a player would be the kind of “lucky” player I alluded to in the opening. League average strikeout rates rend to be around 18.5% (of plate appearances), while average walk rates are around 8.5% of plate appearances. Above average strikeouts rates are obviously those that are less than 18.5% while above average walk rates exceed 8.5%.
Likewise, higher line drive rates suggest a greater likelihood for hits in the future, as line drives are harder to defend than grounders or fly balls. A high percentage of fly balls may suggest more home runs, as well.
Pitching: Fielding Independent Pitching (FIP) – An Alternative to Earned Run Average (ERA)
Similar to batting average, on base percentage, and OPS’ stranglehold of ratio statistics for offense, earned run average (ERA) is likewise vaunted as the supreme barometer for pitching efficiency and success. While it remains descriptive of what actually occurred on the field, it is not the most effective tool for predicting future success. As touched upon at the outset of the article, luck and defensive ability impact a statistic such as ERA. Fielding independent pitching (FIP), however, seeks to isolate the factors that only the pitcher controls to provide a more reliable vision of a pitcher’s abilities and future output. The theory, made popular by research by Voros McCracken in the early 2000s, is that pitchers have little ability to control what happens when a ball is in play (factors that certainly affect ERA). Ultimately, McCracken devised a formula on what a pitcher can control – strikeouts, walks, hit by pitches, and homeruns.
FIP = ((13*HR)+(3*(BB+HBP))-(2*K))/IP + constant
Like wOBA, FIP is put on a scale to compare to a well-established statistic, in this case ERA. Although values change from year to year, the constant value is usually around 3.20 – the difference between the league-average FIP (before the constant) and league-average ERA. FIP value evaluation is the same as ERA interpretation – 4.00 is average, 3.75 above average, 3.25 great, and 2.90 excellent, while 4.20 is below average, 4.50 poor, and 5.00 awful. Pitchers with higher ERAs than FIP scores may be bargains on the market, while converse examples might suggest that it is time to sell high on a pitcher.
Additional Pitching Statistics of Note
Many of the predictive statistics relevant for hitters are also significant for pitchers. Strikeouts and grounders are good, while walks and fly balls are bad. Strikeouts for pitchers are predominately analyzed on a per inning basis as opposed to plate appearances for hitters. The league average strikeouts per nine innings (K/9) value is 7.52. Generally, relievers post higher K/9 values than starting pitchers. While there are inherently many factors involved in such a result, one theory is that starters will face hitters more often and are less likely to fool batters the second or third time through. Of course, starting pitchers are generally more talented, or at least can survive with a greater arsenal, while relievers have fewer pitches mastered. Moving on, league average walk rates are 3.01 per nine, and ground ball rates at 44.7%. Players with weak ERAs but above average K/9, BB/9, and ground ball rates are candidates for stronger future performances and may be bargains on the trade and free agent market.
The Holy Grail of Descriptive Statistics – WAR
Despite categorizing a plethora of sabermetrics as predictive, others are useful descriptive statistics that paint a better portrait than traditional statistics.
For ages, statisticians have attempted to find some way to quantify a player’s all around impact. Wins Above Replacement, affectionately known as WAR, has almost universally been accepted among the sabermetric community as a legitimate “catch all” tool, evaluating offense, defense, and baserunning for hitters, and many of the aforementioned factors for pitchers. WAR is expressed in incremental wins a player produces for his team above what a standard “replacement” player might produce. Again, I must defer to the excellent resources at Fangraphs for a detailed explanation on everything you need to know about WAR and virtually everything about the game. As a huge proponent of WAR, I believe that once again, Mike Trout should once again be the American League MVP, though that is worth a discussion for another day.
Ultimately, sabermetrics provide an advanced view of a timeless game. They are not meant to completely replace traditional statistics, but rather augment player discussion and comparison. Ignoring sabermetrics, however, limits one’s ability to truly understand the game. Likewise, statistical analysis is not meant to eliminate subjective player evaluation, but instead encourages the opposite. Players change and although the statistics highlighted above are meant to aid in projecting the future, it is important to continue to monitor and recognize the growth of physical skill sets and abilities. If baseball analysis was an exact science, there would be no need for debate, and quite frankly, the game would be boring. By combining quantitative and qualitative approaches, however, we are able to make better-informed decisions – not just in sports, but also in all aspects of life. As several franchises such as the Oakland Athletics and Tampa Bay Rays consistently prove, the combination of objective and subjective approaches have allowed them to efficiently produce contenders with fewer resources devoted to the MLB squad. Heavily investing in scouting and development, as those franchises have committed to, is a huge part of the sabermetric process. Today’s reality in the sports business requires the utilization of analytics, and remaining well informed is vital to success.
Georgetown University Class of 2014
Like GSABR on Facebook