Defense, arguably, is the domain of baseball that has experienced the greatest increase in statisticians’ and player-personnel evaluators’ attention in recent decades. As this article from Sports Info Solutions discusses, greater sophistication in assessing MLB players’ defense has helped certain players extend their careers, receive larger contracts, and even get into the Hall of Fame.
There once were only very basic metrics such as putouts (the final step in recording an out, including snaring a fly ball or receiving a throw to first base), assists (e.g., fielding an infield grounder and throwing to first), and errors. Each player’s fielding percentage would then be defined as (putouts + assists) divided by (putouts + assists + errors). Limitations in fielding percentage include its failure to account for plays not involving a putout, assist, or error (e.g., a sharp grounder that a given infielder could not reach but that others at the position might have been able to) and the fact that many players’ fielding percentages converge upon .99 in the long run, even for those not considered especially dexterous with the leather.
Nowadays, advanced metrics based on sky-cam video in MLB ballparks can assess fielders’ range, quickness in reaching the ball, efficiency in routes taken to a fly ball, and so forth. These measurements, combined with more complex mathematical formulas, have accordingly generated more sophisticated fielding statistics than before. For a history of efforts to assess fielding proficiency, see this article from the Society for American Baseball Research (SABR).
The statistic of Defensive Efficiency Record was developed by Bill James roughly 50 years ago to address the question “When a ball is put into play against this team, how often does the defense turn it into an out?” Complexity in the initial formulas for quantifying defensive efficiency prevented the measure from catching on, according to James, but refinements have made it easier to use.
One can think about defensive efficiency in this way. Any team’s fielders will face hundreds of batted balls that they potentially can field – a screaming line drive, a grounder between third and short, a gapper to the warning track, a dribbler in front of home plate, an infield pop-up, and so forth. These examples collectively represent balls in play. The core concept of defensive efficiency is simple – how many of these balls in play does a team’s defensive lineup convert to outs? Note that not all batted balls are considered “in play” as, for example, a home run over the fence is not potentially fieldable.
I like defensive efficiency as a middle-ground metric. Neither as simplistic as fielding percentage nor as complex as modern video-based tools, defensive efficiency is something one can calculate with statistics available on the Internet. Hence, I present defensive-efficiency percentages for the nine baseball-playing Big 12 schools as they head into their conference tournament this week.
By some informal markers, Tech was not as good defensively as the team and their fans might have hoped for, committing more errors (75) than their opponents (56) head-to-head in 56 games. Also, center-fielder Dillon Carter, who certainly covers a lot of ground and converts difficult batted balls into outs, started only half of the Red Raiders’ games due to injury. We’ll see how well these observations translate into the team’s overall defensive efficiency.
As noted by Baseball Prospectus, one can find different formulas for defensive efficiency. Having studied the statistic for many years, I use a formula for which, as I noted, the necessary components are available online (on the Big 12 Conference’s and individual schools’ athletic websites). I’ll spare readers all the details and just review some key points.
We start with the total number of outs a team recorded on defense (multiplying its total innings pitched by 3). We then remove the types of outs that are not due to fielding (e.g., strikeouts, pickoffs, throwing someone out stealing). Because sacrifice bunts and flies are intended by the batter to generate outs, they are also removed. Perhaps more counterintuitively, we also remove double-plays so that the two outs recorded are still credited to one ball in play. We are left at this stage with what might be called defensive outs (outs executed on balls in play).
The final three steps are very easy.
- We first take hits allowed (minus home runs allowed) to represent balls in play that were not converted into outs – grounders that could have been stopped in the infield (and the ball thrown to first for an out) or flies that could have been caught in the air – but were not.
- Second, we add these hits (other than home runs) to defensive outs to obtain the total number of balls in play.
- Third, we divide defensive outs by balls in play, yielding defensive efficiency.
Here are the Big 12 schools, ranked by defensive efficiency in the 2023 regular season.
West Virginia edges out TCU for the top defensive efficiency in the conference, converting 69.3 percent of the balls in play it faced into outs. Texas Tech ranked sixth, with a defensive efficiency of 67.0 percent. Most of the teams are bunched together, so in the Red Raiders’ case, they converted roughly two fewer balls in play per 100 into outs than the Mountaineers. That is not such a large difference, but it could have cost Tech in a close game here or there. (Readers into sabermetrics may realize that 1 – defensive efficiency yields a statistic known as BABIP, opponents’ Batting Average on Balls In Play.)
In comparing different teams, one must assume that they are all exposed to similar distributions of batted balls, that is, similar proportions of grounders, line drives, and flies; similar proportions of sharply and softly hit balls; etc. In other words, one team’s defensive efficiency could conceivably be better than another’s simply because the former had easier balls hit their way. Some teams have better pitchers than others and strong pitching might lead to weaker balls being put into play. Research on MLB data suggests, however, that pitchers have little control over whether batted balls they yield become hits or not.
Note also that fielding errors are typically not part of the formula for defensive efficiency, suggesting that defensive efficiency should be viewed mainly in terms of teams’ ability to get to balls on defense. James initially tried to incorporate errors into defensive efficiency but faced difficulties because “not all errors put runners on base” (e.g., a ball rolling to an outfielder for a presumed single being fumbled into a double, where the runner would have gotten on base anyway). James’s attempts to parse different types of errors (and other issues) increased the formula’s complexity, so were abandoned. One defensive-efficiency formula I came across included when batters reached on errors (ROE) but because ROE was not available on the sources I consulted, I did not use this version of the formula.
Defensive efficiency is not a perfect measure, resting on certain assumptions and approximations. However, unless a team has access to sky-cam video technology and the staff resources to process all the data on fielders’ travel distances in response to balls hit in their vicinity, defensive efficiency may be the most practical option. It is certainly more valuable than fielding percentage. By studying a team’s defensive efficiency at various points in a season, one can see how well the team – whether through shifting, quickness, or fielding technique – is preventing its opponents from, in the immortal words of Wee Willie Keeler, “hit[ting] ’em where they ain’t.”
Alan Reifman is Professor of Human Development and Family Sciences at Texas Tech University
