The Evolving Understanding of Baseball Sabermetrics

by Hollywood Sports

While alternating between watching Game Six of the Stanley Cup Playoffs and Sunday Night Baseball between the Los Angeles Dodgers and the Atlanta Braves, ESPN displayed a surprising list: Pitching Leaders in BABIP. 

BABIP is an acronym for “Batted Average on Balls In Play”. Removing home runs and strikeouts from a hitter's batting average is a more specific manner to detail batting average perhaps offering better insight into how hitters and pitchers are performing. It is an offshoot from some of the ideas initially espoused by early sabermetrician Voras McCracken who developed the influential metric “Defensive Independent Pitching Statistics.” I mention McCracken because his seminal paper on this subject offered a narrative of his as a child hitting balls in a cemetery where he discovered that what happens to the batted-ball put into play is purely a function of luck. 

What struck me at the time was how mediocre of a wiffleball hitter McCracken must have been. Any backyard wiffleball hitter developed at least some skill in being able to place a batted ball in a certain direction. But it had become a foundational established truth of sabermetrics that hitters had no control over the batted ball, despite the seeming inconsistency that hitters had plenty of control over the batted ball if it landed over the fence for a home run. The fantasy baseball “experts” at ESPN would eager to deploy BABIP to identify players who were due for improved or declining numbers based on their BABIP number. Pitchers with high BABIPs were unlucky and those pitchers with low BABIPs were too lucky and sitting ducks to get blown up in their next appearance. Using the same logic, hitters with high BABIPs were riding good fortune, and hitters with low BABIPs were likely due to see better numbers. The fact that future Hall of Famers like Ichiro Suzuki was usually at the top of the BABIP list for hitters and Clayton Kershaw was at the top of the lowest BABIP numbers for pitchers gave little direct pause to the experts (probably not enough space at the website to detail that along with the trials and tribulations of their fifth-place teams in all their “experts” league in content as compelling as food pics on Twitter). 

In later years, these fantasy baseball experts (a few who are still detailing their latest experts league exploits even today) reigned in their BABIP as “all luck” to then claim it was all the amateur fantasy baseball players who were misusing BABIP. The expert manner to deploy the metric was more nuanced and sophisticated, you peons, yet this must have been ESPN's proprietary information as these experts rarely communicated how to evaluate the nuances. And the assumption remained that Ted Williams simply had no power to impact where the ball went off his bat. 

So to now see ESPN now fully incorporate BABIP into their “endorsed” statistics for their Sunday night broadcast — but with the assumption that BABIP is a component of skill — well, I was amused. This is the same network that will have their 4th-and-1 percentages come down from the heavens in the fall to tell you how much smarter they are than the idiots coaching football teams — all setting up their 12 hours of Hot Take Debate TV programming the next day. 

Fielding independent statistics was always more complicated to discover hidden meaning that McCracken and ESPN fantasy baseball experts suggested. But to paraphrase Marshall McLuhan, the “medium is the message” — and in this instance, the appearance of having more sophisticated knowledge of baseball is too often the point of the message, regardless of the coherence of the argument. There is a reason why these experts get to selectively mention the current state of their dozen fantasy baseball teams. They are experts because we were just told they are experts. 

For handicappers and bettors, having the “expert” business card laminated for posterity does not help at all if picking loser after loser at the betting window. For me, I experimented for a few years comparing team BABIP numbers versus the starting pitcher’s BABIP to determine discrepancies: playing behind the same defense, were certain starting pitchers generating big gaps in the net BABIP differential? If so, that difference could help identify starting pitchers who were experiencing luck/unluck relative to their baseline numbers. What I am looking for is underlying numbers that offer conflicting evidence to how a starting pitcher looks when only evaluating Win/Loss record, ERA, and WHIP. Eventually, I began analyzing these discrepancies between team BABIP and starting pitcher BABIP but made it specific to ground balls and line drives. A batter probably does not want to hit a ground ball, so those numbers might illuminate luck (or unluck for the pitcher) if the BABIP for ground balls is high. On the other hand, hitters want to hit line drives — so high Line Drive BABIPs is probably a good thing for hitters (and a bad thing for pitchers).

I had pretty good success with this approach, but it took a lot of work. Eventually, I concluded that the advanced ERA formulas used by SIERA and xFIP did a fine enough job of identifying starting pitchers that were due a visit from the Regression Gods, either in a good or bad way. The xFIP metric stands for Expected Fielding Independent Pitching (admittedly, from the McCracken family tree of logic). I like it because it normalizes the home runs that a pitcher allows from their fly ball rate since home runs allowed are also dependent on other factors like the ballpark). The folks at Baseball Prospectus developed SIERA (Skill Interactive Earned Run Average) in a complicated formula that attempts to project the ERA moving forward by taking into account the types of batted balls a pitcher puts into play. I like this tinker because it leaves room for the possibility that hard-hit balls (and direction) are a function of skill. It reminded me of the work I was doing with BABIP specific to Line Drives (hard-hit balls). 

These days, I use xFIP and SIERA as guides — but it is only one of the tools in my toolboxes. Most pitching metrics have a bias against ground ball pitchers despite the sustained success of pitching coaches like Dave Duncan of the St. Louis Cardinals who seemingly had his pitching staff overachieve for years. Fortunately for me, my reliance on team trends kept me away from betting against too many Cardinals’ starting pitchers hoping that their poor fielding independent numbers (like low strikeouts) would finally start catching up with them. 

Sports analytics continues to evolve, especially for those of us dependent on good results rather than self-serving propaganda. Look for basketball and soccer to improve in their evaluation of players and teams when they start detailing expected baskets/goals versus an individual’s net edge versus expected baskets and goals. Like Ichiro, Kershaw, and Ted Williams, Stephen Curry and Lionel Messi are elite players for a reason. 

Best of luck — Frank.

All photographic images used for editorial content have been licensed from the Associated Press.

© 2022 Al McMordie's BigAl.com. All Rights Reserved.