Baseball’s Advanced Statistics

Everyone in the world knows that a .300 batting average is good on the day they’re born (or, at least they should). But what about other lesser known statistics in baseball that seem to be gaining ground by the day? You’ll see me using advanced statistics such as on-base percentage (OBP), slugging percentage (SLG), on-base plus slugging (OPS), weighted on-base average (wOBA), ultimate zone rating (UZR) and Fielding Independent Pitching (FIP) a lot on this blog. So what do they mean? A primer of sorts is below.

For reference in the primer: AB = at-bats, PA = plate appearances, H = hits, HR = home run, BB = bases on balls (walks), HBP = hit by pitch, TB = total bases, NIBB = non-intentional walks, RBOE = reached base on error.

On-base percentage (OBP): (H + BB + HBP) / (PA). Reaching base safely is the best thing you can do as a hitter, whereas making an out is the worst thing you can do. On-base percentage is thereby a very simple, yet incredibly important statistic that’s often overlooked by fans. On-base percentage is the frequency in which a hitter does not make an out, making it the single most important isolated offensive statistic in baseball.

On-base percentage is a much truer indication of a hitter than batting average because on-base percentage takes all plate appearances into account rather than just at-bats. On-base percentage examines every time you step into the batter’s box, not just selected balls in play and strikeouts (like batting average), so on-base percentage is a much more complete view of a hitter. Most important regarding the distinction between plate appearances and at-bats is that plate appearances take walks into consideration, because as everyone’s little league coach told them when they were eight years old, a walk can be just as good as a hit.

Batting average and on-base percentage can give off completely different indications on a player — for example, a .300 hitter who hardly ever walks could have a .315 on-base percentage (below average). As a general rule, a .315 or below on-base percentage is poor; .340, league average; .370, good; .400 or above, fantastic. On-base percentage is best indicator of a hitter’s production save for weighted on-base average (wOBA). The single-season leader in on-base percentage is Barry Bonds at .6094 in 2004. Ted Williams, with his legendary batting eye, is the career on-base percentage leader at .482.

Slugging percentage (SLG): (TB) / (AB). Often paired with on-base percentage to form OPS (on-base plus slugging), slugging percentage is a less indicative measure of a hitter than on-base percentage. Nevertheless, slugging percentage is an effective yet simple measure of how well a player hits for power. A single is worth one base; a double, two bases; a triple, three bases; a homer, four bases. Divide the total number of bases by at-bats and you have slugging percentage.  The more homers, triples and doubles a batter hits, the more total bases a batter attains, and the higher his slugging percentage will be. A slap singles hitter probably won’t have a high slugging percentage.

Slugging percentage also seems a little faulty because even though it’s measuring a player’s power, two of its factors — doubles and triples — can be more indicative at times of a player’s speed than his power. In general, a .390 or below slugging percentage is poor; .420, league average; .460 or above, good; .500 or above, great. A perfect slugging percentage is 4.000 because it means every at-bat is resulting in a home run. The single-season sluggling percentage leader is Bonds at .8634 in 2001. The career slugging percentage leader is Babe Ruth at .690.

Isolated Power (ISO): (SLG) – (AVG). By subtracting a player’s batting average from his slugging percentage, we’re isolating the doubles, triples and homers that make up slugging percentage. In other words, slugging percentage minus singles. The more extra base hits, the higher the isolated power.

On base-plus slugging (OPS): (OBP) + (SLG). This is great for a quick look at how productive a hitter is, as it combines how well a hitter gets on base and hits for extra bases. Sometimes, it seems like it could be the be-all, end-all for hitting – it’s great for a quick look at the productivity of a hitter. If a hitter’s OPS is .875 or above, they’re really good. If a hitter’s OPS is down around .725 or lower, they’re not very good, with .760 being about league average.

The single-season OPS leader is Bonds at 1.4217 in 2004. The career OPS leader is Ruth at 1.164. Still, though, there are some problems with OPS, largely because it undervalues on-base percentage — it’s generally accepted that on-base percentage is a better indication of a hitter than slugging percentage, but slugging percentage usually makes up a larger chunk of OPS than on-base percentage.

Also, adding on-base percentage and slugging percentage, which have two different denominators — one uses plate appearances; the other, at-bats. We learn in third grade not to combine two different denominators. In addition, the positive outcomes from hitting aren’t weighed properly with OPS – a homer and a swinging bunt count the same towards a hitter’s on-base percentage. Measuring the productivity of a single and a homer through the use of total bases like slugging percentage also isn’t proper. These issues with OPS are solved with the creation of…

Weighted on-base average (wOBA): ((0.72 x NIBB) + (0.75 x HBP) + (0.90 x 1B) + (0.92 x RBOE) + (1.24 x 2B) + (1.56 x 3B) + (1.95 x HR)) / (PA). The coefficients are pieces of data pulled from analysis showing exactly how much value there is for each positive outcome for a batter. Every positive outcome for a hitter is weighed properly, unlike on-base percentage, where a homer counts the same as a walk, and also unlike slugging percentage, because total bases isn’t a proper way to weigh the value of, say, a double or a homer.

Also, wOBA is conveniently scaled to look like on-base percentage, so .315 or below is poor, .340 is league average, .370 is good, .400 or above is fantastic. wOBA is now considered the best singular statistic out there to measure how productive a hitter is. The career wOBA leader is Ruth at .510. Click here for more on wOBA, courtesy of Yahoo! Sports’ “Big League Stew.”

Batting average on balls in play (BABIP): The rate at which balls in play (not including home runs) fall in for hits. For both pitchers and hitters, rates tend to hover around .300, the usual rate that a batted ball falls in for a hit.  Hitters’ rates can seesaw a little more than pitchers’ rates. A hitter’s career BABIP can be a better indicator as to what a hitter’s BABIP should be during a given season moreso than the .300 mark. That .300 BABIP is steadier for pitchers.

For hitters, a BABIP well above .300 probably means they’ve been a bit lucky and should even out over the course of a six-month season, while a BABIP well below .300 means they’ve been unlucky and should see an uptick in production down the line. But if a hitter’s BABIP is really low, it could just mean they can’t hit, especially for cup-of-coffee type of players who don’t last long.

For pitchers, when a BABIP is well above .300 in a given season, it’s usually a sign that the pitcher has been unlucky, which should even out by the end of the season and move towards the .300 mark. (Of course, if a pitcher’s BABIP is consistently well above .300, it could just mean that they’re terrible.) When a pitcher’s BABIP is well lower than .300, they’re due for more balls to fall in against them in the long run. Click here for more on BABIP, courtesy of Yahoo! Sports’ “Big League Stew.”

Contact Percentage (Contact%): The rate at which a hitter makes contact with a pitch that he swings at. League averages tend to hover a little above 80 percent.

Outside Swing Percentage (O-Swing%): The rate at which a hitter is swinging at pitches outside the strike zone.

First Pitch Strike Percentage (F-Strike%): The rate at which a pitcher is throwing the first pitch of a given hitter’s plate appearance for a strike.

Swinging Strike Percentage (SwStr%): The percentage of all pitches a batter sees that result in swings and misses. League averages are about 8.0 percent.

Fielding Independent Pitching (FIP): (13HR + 3BB – 2K) / IP. (Add 3.20 to scale the raw FIP number to look like ERA.) The theory behind FIP is that pitchers can control three things — home runs allowed, walks and strikeouts. For everything else, the theory goes, whether batted ball becomes a hit or not can be largely based on randomness. In other words, Pitcher A doesn’t control whether balls in play become hits any better than Pitcher Z. Pitcher A, though, can control strikeouts, walks and homers allowed better than Pitcher Z.

By measuring a pitcher’s homers allowed, walks and strikeouts, it can tell us how well a pitcher controlled the aspects of a game he can control, and not so much how lucky or unlucky he got. Obviously, the more strikeouts a pitcher attains and less walks and homers he gives up, the lower his FIP will be. While ERA tells us how many earned runs per nine innings a pitcher is giving up, FIP tells us how well a pitcher is pitching regardless of luck, and FIP is thus a much better idicator of future performance for a pitcher.

If there is a significant difference between a pitcher’s ERA and FIP, luck will probably even itself out and a pitcher’s ERA will become much closer to FIP over the course of a six-month season. FIP is scaled to look just like ERA, so anything below 3.00 is great, anything from 3.00 to 3.75 is good, and anything above 4.50 is poor. FIP can be unreasonably kind to pitchers who are getting really lucky with their home run to fly ball ratio, why is why there’s also xFIP (expected FIP). xFIP takes home run-to-fly ball ratio into account to formulate how many homers a pitcher will probably give up over the long haul. The career FIP leader is Rube Waddell at 1.92 of the old St. Louis Browns (that would be the Baltimore Orioles’ franchise, for your information). Click here to learn more about FIP.

Ultimate Zone Rating (UZR): The amount of runs a particular player is saving his team in the field at his position in comparison to an average fielder at that position. A 0.00 UZR is average, so any mark above that means a fielder is better than an average fielder at his position, whereas a negative UZR means a fielder is a sub-average fielder. It’s best to look at UZR over the course of at least three seasons and not just one season. UZR’s calculations involve splitting the field into 64 “zones.”

Of course, UZR is a bit “theoretical,” so to speak — if Player X has a UZR of 5.00 for a season playing right field, there’s no way of definitely knowing that Player X saved his team five runs over what an average right fielder would have done in that same season. But all statistics and metrics have limitations, and there is rational reasoning behind the UZR statistic. With UZR and other fielding metrics, we now have a pretty good idea who the good defenders are and who the bad ones are. There is also UZR/150, which is UZR pro-rated over the typical 150 games a player usually plays over the course of a 162-game season. Learn more about UZR here courtesy of Yahoo! Sports’ “Big League Stew.”

Wins Above Replacement (WAR): The amount of wins that can be attributed to an individual player due to past performance over that of a replacement player. A replacement level player is basically the Triple-A players that can get sent up and down from the minors — a freely available, run-of-the-mill Quadruple-A player, basically. WAR measures everything — hitting and defense for a position player — all within the framework of their league and position in a particular season. (For example, replacement level production at first base is going to be significantly higher than for a catcher.) For every 10 runs a player is above replacement, he’s individually worth one win to their team.

A replacement level player is a 0.00 WAR, so if Player X has a WAR of 2.00, he’s two wins above replacement. A fantastic WAR is 6.00 or above, 4.00 or above is good, 2.50 is about league average, and a negative WAR means that player is theoretically worse than replacement-level player. The sense I get is that some fans think picking up a good free agent adds 10-12 wins and an injury to a star player can cost a team 10 wins. That’s not true unless you’re adding or subtracting Barry Bonds circa 2001-2004 or a very highly performing MVP or Cy Young Award-winner type. 2009 AL Cy Young Award winner Zack Greinke, for example, led the majors in WAR with 9.4.

WAR is very much an imperfect statistic — there’s nothing to say that Player X is definitively two wins above replacement, whereas one knows that if a hitter has a .250 batting average, he’s definitely getting a hit once every four at-bats. The concept of a “replacement level player” at a given position is rather abstract and ever-changing, while looking at the number of wins above replacement is also abstract. What’s a 4.1 WAR player? In other words, what’s a tenth of a win? But these aren’t reasons to disregard WAR — all metrics and statistics can be said to have problems with them.

The defensive aspect of WAR for position players is especially difficult to quantify because there’s no perfect defensive statistic. UZR is the go-to defensive statistic these days and is what FanGraphs uses as their defensive component in their computation of their version of WAR, but who’s to say UZR is the best defensive statistic? There is no way to know just how many runs a defender has saved. Generally, good defenders can be overrated by WAR, and bad defenders, underrated. For example, say Player Y has been rated as a player who costs his team runs defensively for years now. But one year, Player Y rates well defensively, but it looks like an obvious outlier, and his WAR shoots up. Take these kinds of one-year spikes in WAR driven by defense with a grain of salt.

There are also still different ways to calculate this all-encompassing statistic — Baseball Reference (brWAR) and FanGraphs (fWAR) are two sites that have come up with a WAR statistic. I’ll mostly be using fWAR on this blog. WAR also is good for finding the value of a particular player in comparison to his salary — since the market paid about $4.5 million for a win in 2009, the salary of Player X should theoretically be about $9 million. Click here for more on WAR, courtesy of Yahoo! Sports’ “Big League Stew.” Daniel Moroz of Camden Crazies explains how to actually calculate WAR here. FanGraphs has an explanation of their version of WAR here.