1/8/13

Michael Friere - What Are Sabermetrics - BABIP


Last week, part one of this series of articles started with an introduction, followed by a discussion centered around a fairly common pair of statistics called OPS and OPS+ (which are more valuable then the older standard of batting average).
This week I want to look closer at a statistic called BABIP, also known as batting average on balls in play. I am being a bit sneaky by using this statistic, since it can be used to evaluate a batter’s performance, as well as a pitcher’s performance, which makes it quite versatile.

In analyzing OPS and OPS+, we inadvertently discussed batting average, since it is a component of the listed statistical formulas. Batting average (AVG) is basically the number of hits, divided by the number of official at bats (which does not take other things into account, such as reaching base via a walk).
So, when discussing BABIP, a basic understanding of AVG is essential.
Fine, so what the heck is BABIP, why should I care about it and and how does it apply to BOTH batters and hitters?
A more detailed definition would be the percentage of plate appearances ending with a batted ball in play, that is scored a hit (with the exception of home runs which cannot be fielded). For a batter, it is person specific, whereas a pitcher’s BABIP is related to the total number of hitters he has faced.
OK, so how do we actually figure out how to calculate BABIP? The general formula for BABIP = (hits - home runs) divided by (total at bats - strikeouts - home runs + sacrifice flies). The result of the listed calculation should be expressed similar to batting average, i.e. Player A had a BABIP of .325 for 2011.
To make further sense of the statistic, a BABIP around .300 is considered the average, or the norm for the player in question. That number can fluctuate from player to player, or even season to season. Certain factors, such as the ballpark (think Coors Field versus Citi Field) or if the pitcher is a ground ball or fly ball pitcher, can have a modest effect on the average. However, a BABIP of .300 is considered a benchmark, sort of like 100 is the average for OPS+
Knowing how to generate the BABIP and what is considered a “good” BABIP is fine. You won’t hear opposing fans arguing over who has a better BABIP, necessarily and most casual observers won’t even know what it is.
But, you should care because it is an excellent statistic to see if a specific player is overachieving or underachieving for a specific period of time. That is extremely helpful when general managers are trying to figure out who to sign and who to pass on, whether it is your own free agent, or a player from another team that is now available.
Plus, we all know that Sandy and Co are statistically inclined and they absolutely love finding the diamond in the rough, i.e. the undervalued asset. BABIP is a valuable tool when trying to assess a player’s current performance and what they may do in the near future.
Here is scenario for you to consider. Player A, a free agent short stop, has been in the major leagues for seven seasons and has averaged a BABIP of .305 for the first six years of his career. The player had what some call a “career year” or “contract year” in year seven and produced elevated statistics with regards to batting average, on base percentage and OPS (all directly or indirectly influenced by BABIP).
Further analysis shows a BABIP in year seven of .365, or sixty “points” higher then the previous six year career average. Odds are Player A simply benefitted from good fortune in year seven and his statistics will drop sharply in the future, when the BABIP regresses to the mean (returns to the previously established average).
Paying that player for the past year, or expecting that sort of performance in the future is unlikely. As a matter of fact, that is a perfect candidate for a future label of “free agent bust” because they overachieved and they are overvalued.
How about another scenario? Player B, a free agent pitcher, has a ten year track record of success, to include an average BABIP of .295 over that time. The past two seasons, the pitcher’s statistics and overall performance have dropped suddenly. Ruling out age and injury factors, you see that the pitcher’s BABIP over the past two years averaged .345, which is fifty “points” above the previously established average.
Could an increase in BABIP of fifty “points” have a negative effect on a pitcher’s stats? Yes and it is an often overlooked statistic.
In that vein, Player B is a good candidate to “bounce back” in the coming years since it makes more sense to look at ten years of consistently low BABIP, versus a short term increase. When that player’s BABIP regresses to the mean, the corresponding statistics will also improve. That player has underachieved and is undervalued. Just the sort of player that Sandy would be on the prowl for.
Keep in mind, BABIP is only one tool in the evaluation process. It cannot forecast things like age, injuries and unrelated improvement. In the first scenario, maybe Player A dedicated himself to a fitness program and his performance improved as a result? Or, in the second scenario, maybe our undervalued asset is simply getting old and his overall performance is suffering as a result.
While it does have shortcomings, BABIP is a neat statistic and one that I am sure is being used in our very own front office. You can use it to make your own assessments on players the Mets may be interested in, or current Mets’ players who are on the brink of potentially being replaced.
Since I am sure some fans are curious, Player A listed above is not Jose Reyes, rather an imaginary example I made up. However, since some of you are curious, Jose Reyes has a career BABIP of approximately .314 (helped in some part by his insane speed leading to extra hits on balls that would be outs for slower players). In 2011, Jose had a BABIP of .357 (and that is after slumping a bit in the second half of the year), which is approximately .043 “points” above his average. Think that had anything to do with his batting title and .337 average?
If Jose regresses in the future, and I think that is likely since an increase of .043 is unsustainable, future BABIP’s in line with his career average would still yield a solid batting average in the upper .290’s, with a corresponding drop in the other related statistics. Very good player? Yes, of course. Overvalued, injury prone and getting older (especially for a player who relies on his legs so much)? Sadly, yes. Worth a long term, nine figure contract? Not if I am in charge of pulling the trigger.
Feel free to use BABIP for other Mets players, such as Angel Pagan, David Wright, Mike Pelfrey, etc. It may help shed some light on what will be an interesting offseason of player movement.


No comments: