How lucky or unlucky were the 2014 St. Louis Cardinals on balls in play?

Batting Average on Balls In Play (BABIP) is a stat that measures exactly what its name indicates. It looks at how often a batter notched a hit when putting the ball in play. Having the denominator of balls in play means that BABIP has a narrower focus than batting average, on-base percentage, slugging percentage, OPS, or wOBA.

What constitutes a ball in play? A home run is not considered a ball in play. Neither is a strikeout or a walk. A hit by pitch isn't a ball in play either. The following events are considered balls in play: singles, doubles, triples, groundouts, fly outs, fielder's choices, reached on error, and sacrifice hits (whether they be bunts or flies).

Here's the formula for calculating BABIP:

( H - HR ) / ( AB - SO - HR + SF)

MLB players collectively typically post a BABIP around .300. Last year, the MLB-wide BABIP was .299. Over the last ten years, the BABIP for all big-leaguers has ranged between .295 and .303.

The rule of thumb is that a player with a BABIP below .300 has been unfortunate on balls in play and one with a BABIP over .300 lucky. The problem with this rule is that it is premised on treating every ball in play the same, regardless of batted-ball type or how hard it was hit. This makes the .300 principle rather flawed.

Because MLB will not make its Hit F/X data public, we have no way of knowing exactly how fast the ball flies off the bat of a player. That being said, I feel safe stating that the average ball off Matt Holliday's bat has more velocity to it than the average ball off Daniel Descalso's bat. Holliday hits pitchers harder than Descalo. It seems safe to surmise that the greater a batted ball's speed, the more likely it is to result in a hit. So player's that make better contact ought to have a higher BABIP than average.

This is backed up by the data. As a general rule, a line drive is more likely to fall safely for a hit than a grounder or a fly ball. That's because the liner is typically the hardest hit of the batted ball types. A groundball is more likely to be a safe hit than a fly ball. In the BABIP hierarchy, it's: LD > GB > FB. (It's worth reminding ourselves here that homers are not balls in play and not included in BABIP's calculation. Most homers are fly balls. So fly balls have more value when calculating broader offensive stats like BA, OBP, SLG, OPS, or wOBA than they do BABIP.)

This is the other problem with the .300 BABIP rule of thumb. The more liners a batter strikes, the higher his BABIP is likely to be. Put otherwise: a player with an above-average LD% is likely to have an above-average BABIP. This is a reflection of his skill at barreling the baseball and making good contact. Likewise, the faster a player is, the higher his BABIP is likely to be. There is undeniably an element (or two or three) of skill to BABIP, even though the stat is extremely volatile and fluctuates wildly from season to season.

The fluctuation of BABIP and the effect that park and defense can have on the stat has led folks to explore how best to calculate Expected BABIP (xBABIP). VEB community member Paulspike developed such a formula and was kind enough to share his spreadsheet with me. It's based on the following: LD%, GB%, IFH%, Fangraphs Speed (Spd), Isolated Power (ISO) and Contact%. These inputs attempt to capture a player's skill for making good contact and his ability to turn a batted ball into a hit with his feet.

I put the 2014 St. Louis Cardinals through Paulspike's xBABIP calculator. Then, using the work of Jeffrey Gross at The Hardball Times, I calculated each player's xBA, xOBP, xSLG, xOPS, and xwOBA. Gross lays out the problems in doing this and the decisions one has to make. I elected to count every hit (whether added or subtracted due to his xBABIP) as a single. I did so because I wanted to be conservative; I didn't want to give a player too much credit or subtract too much from his production by bringing extra-base hits into the equation.

These stats are mere approximations of what a player's 2014 might have looked like with a BABIP closer to what his peripherals indicate was possible. Ingest a grain or two of salt when looking at the numbers. Further, keep in mind that BABIP has very little predictive value and neither does xBABIP. Nothing in this post gives much of an indication (if any) of what to expect from a player in 2015.

Without further ado, the charts:

Matt Carpenter

Yadier Molina

Matt Holliday

Matt Adams

Kolten Wong

Jhonny Peralta

Jon Jay

Peter Bourjos

Tony Cruz

Allen Craig

Daniel Descalso

Mark Ellis