Advanced statistics: How to use BABIP or batting average on balls in play

Fanalytics, Sabermetrics and all advanced baseball statistics have been created to help us better understand the game. Many of them were created by fans and fantasy baseball managers to help predict future performance and, in return, win fantasy baseball.

Robert McIlhatton sent an email to the Chinstrap Ninjas recently questioning one of the statistics, batting average on balls in play or BABIP, as a barometer for luck. The email prompted this article, but we’ll get into the specifics of his email in a later post.

First, let’s get as up to speed as I can take you on BABIP, and how it can help you win fantasy baseball.

Let’s get this out of the way: For many Chinstrap Ninjas readers, this is not the article they were looking for. If you have already overturned a number of stones to find the dark secrets of BABIP, you might become bored and drift off. You are in charge of pointing out the many typos and errors that are sure to follow.

Let’s get this out of the way, Part II: I am not Bill James or Ron Shandler or any of these other icons of baseball statistickery. I did not invent the gauges, but the concepts are interesting enough and helpful enough for me to have done quite a bit of research.

BABIP Basics

BABIP means Batting Average on Balls In Play. Most baseball fans know batting average is hits divided by at-bats. BABIP takes the statistic one step further, cutting out any at-bat that does not end with the batter putting a ball in play.

The league average BABIP is around .300. That means 30% of the time a hitter can expect to get a hit when they put the ball in play. Defenders (and pitchers) can expect to get an out 70% of the time a ball is put into play.

The formula can be used to help predict luck, or a correction of said luck, for both batter and pitcher performances. Batters are a little more complicated, so let’s dig in there first.

BABIP and batters

We can compare any player’s BABIP to the baseline (.300), but many players will establish personal norms slightly above or slightly below .300. Line-drive hitters, like Shin-Soo Choo and Josh Hamilton, and faster players, like Ichiro Suzuki and Michael Bourn will have naturally higher average¬† BABIPs in the .330-.350 range.

A BABIP decline means defenders and pitchers got a little more “lucky” against hitters, snagging a liner or stealing away a home run. Put luck in quotes because we’re talking about a combination of¬† luck AND the oppositions’ defensive/pitching talent. A higher BABIP means the hitter got “lucky” and dropped in a few more bloopers, seeing-eye singles or scooted one just past the third baseman.

Declining/improving skills — whether because of in-season pitching/batting adjustments or more long-term changes because of age, severe injury or skill development/maturation — can also cause small changes in BABIP from year to year.

But when a player fluctuates significantly from their norm we can expect their BABIP to correct the following year and take their batting average along with it. You could assume that as “luck” returns, so could a few runs, homers, steals, etc., but that is not usually the case. Opportunity (at-bats, at bats with runners in scoring position, etc.) plays a bigger role in those categories.

It should be pretty obvious how we can use BABIP to determine batting average sleepers and busts in fantasy baseball, generally speaking. But I’m sure you inquisitive ninjas have begun to crank out questions. But what about good hitters vs. bad hitters? Sluggers vs. speedsters? Here are some examples of 5-year BABIP numbers that show the statistic is telling across various player types:

Batting/slugging All-Stars

  • Albert Pujols:
  • Five-year average BABIP: .312, AVG: .330
  • Best BABIP: .350 (2008, AVG that year: .357)
  • Worst BABIP: .290 (2006, AVG that year: .331)

Hanley Ramirez has established himself at a higher-than-usual BABIP, but it’s hard to argue against Pujols as a premiere slugger. He has a surprisingly average BABIP. Also note how the batting average rises and falls in concert with his BABIPs.

All or nothing sluggers (high K rate):

  • Adam Dunn:
  • Five-year average BABIP: .302 Average: .252
  • Best BABIP: .330 (2009, AVG: .267)
  • Worst BABIP: .260 (2008, AVG: .236)
  • Mark Reynolds:
  • Three-year average BABIP: .310 Average: .232
  • Best BABIP: .340 (2009, AVG: .260)
  • Worst BABIP: .260 (2010, AVG: .198)

Again we see averages following along with BABIP. You can see by looking at Reynolds BABIP in 2010 why some people are calling him a sleeper in 2011. If his BABIP/average correct back toward his norm, he’s going to hit .230 or .240 which is palatable considering the other numbers he produces. Again, note how average the average BABIPs are.

Slap-hitters/speedsters

  • Ichiro Suzuki:
  • Five-year average BABIP: .362 Average: .330
  • Best BABIP: .390 (2007, AVG: .351)
  • worst BABIP: .340 (2008, AVG: .310)

It shouldn’t surprise you that Ichiro has such a ridiculous average BABIP, but look at the difference when his BABIP drops.

We might expect free swingers with high contact rates to be different because they put a lot more balls in play. We would be wrong:

High contact rate (high and low batting averages)

  • Vladimir Guerrero:
  • Five-year average BABIP: .316 Average: .310
  • Best BABIP: .330 (2006, AVG that year: .329)
  • Worst BABIP: .300 (2010, AVG: .300)
  • Jimmy Rollins:
  • Five-year average BABIP: .274 Average: .269
  • Best BABIP: .300 (2007, AVG that year: .296)
  • Worst BABIP: .250 (2010, AVG: .243)

Both Guerrero (85-90% last five years) and Rollins (88-91%) have exceptional contact rates, and yet, despite putting more balls in play than many of their major league colleagues, their BABIPs are not only good reflections of their batting averages, they mirror each other closely.

As with any statistic, there are exceptions and outliers we can point to, but as these examples show, across all player types BABIP is as good a tool as we have to predict “lucky” or “unlucky” batting averages.

Pitcher predictions with BABIP

The principles explained above all fit into the pitcher discussion as well. We just have to flip the numbers around. Pitchers (and their defenses) that allow more than 30% of balls in play to drop are suffering from “bad luck.” Pitchers who have a low BABIP have “good luck.”

Pitchers, however, are more likely to correct toward the .300 norm than hitters. The difference is fairly simple:

A batter has 4 to 6 chances to put a ball in play during each game. Starting pitchers could face up to 30 or 40 chances for a ball to be put into play. More samples — that include a mix of average, above-average and below-average batters — means the numbers will be closer to average.

Shutdown relievers like Mariano Rivera, who typically posts BABIPs in the .230-.250 range and human rollercoasters like his teammate, Joba Chamberlain, who has averaged a .330 BABIP over the last three years, represent reliever BABIPs which fluctuate more based on player skill.

Your starting pitcher examples are two of the most consistently awesome pitchers in baseball the last three years:

  • Roy Halladay: 2010 BABIP: .300, 2009 BABIP: .310, 2008 BABIP: .290
  • Tim Lincecum: 2010 BABIP: .320, 2009 BABIP: .300, 2008 BABIP: .310

Note that Lincecum put up the worst ratios of his career last season and had a higher than average BABIP. However, his best season (high Ks, best ratios, lowest BB/9) was in the season he had an “average” BABIP.

You say, “but what about pitchers who rely less on dominating hitters with strikeouts and more on letting batters put the ball in play?” I present to you, Mr. Never-Get-A-Strikeout:

  • Mark Buehrle: 2010 BABIP: .320, 2009 BABIP: .280, 2008 BABIP: .310

Average those BABIPs over three years and you get .303.

Not only are BABIPs a more complete predictor for pitchers, they also usually follow the true .290-.310 average BABIP.

It’s worth repeating that when players show improvement or decline in their underlying skills, it will also have minor affects on BABIP from year to year. However, those indicators are all skill based.

BABIP is the one statistic that can come close to allowing us to gauge something as unquantifiable as luck.

Comments are currently closed.