/cdn.vox-cdn.com/uploads/chorus_image/image/55449463/674112892.0.jpg)
Defensive metrics have always been the weakest link in statistical analysis of players’ performances in Major League Baseball. The newest defensive metric for outfielders is “catch probability.” The numbers in this metric show more accurately how a player is performing against the league average percentages when it comes to making plays on fly balls.
Along the facing of the upper deck near home plate at Comerica Park, and at ballparks around the major leagues, there is affixed a large black box that looks somewhat like a loudspeaker. It is actually the Statcast box, which uses radar to measure things that were not previously measured in baseball metrics.
Revised Zone Rating (RZR) measures the percentage of balls hit into a predefined zone that are turned into outs at each position. The “zone” at each position is defined as the area where at least 50 percent of plays are converted to outs, without regard to the speed or trajectory of the ball, or positioning of the fielder. Balls outside of that zone that are converted are recorded at “out of zone” (OOZ) plays.
Statcast measures exit velocity- which is the speed of the ball off the bat; launch angle- which is the trajectory of the ball; hang time, and the positioning of players at the start of a play. As of May 2017, Statcast also factored in the direction that the fielder has to travel, factoring in when a fielder has to go in the opposite direction of home plate to make a play.
Zone rating was subject to random variation swings due to the number of hard hit balls and adjustment of where fielders are positioned. Statcast eliminates most of that. While we only have small samples of Statcast data for catch probability, these metrics will be much more reliable than even larger samples that do not factor in such critical variables.
Catch probability is the percentage of time that an outfielder converts a fly ball into an out using all of these measurements, based on how other fielders have performed given the same amount of time and distance to make a play. A zero percent catch probability is a ball that no fielder makes, ever. A 100 percent catch probability is a play that is made every time.
Catch probability is presented in five categories, based on the percentage of time that a particular ball would be caught based on real data collected from all balls put in play in the major leagues. A one star play is one that would be converted 91 to 100 percent of the time, based on the given variables. A two star play would be made 76 to 90 percent of the time, and a five star play would be converted 0 to 25 percent of the time.
To look at how these numbers work, we will look at the Detroit Tigers’ outfielders catch probability statistics for the three months of the 2017 season. Baseball Savant has the full breakdown of all mlb outfielders here.
2017 Detroit Tigers’ Outfielders’ Catch Probability
Player | 5 star plays/ opps/ pct | 4 star plays/ opps/ pct | 3 star plays/ opps/ pct | 2 star plays/ opps/ pct | 1 star plays/ opps/ pct |
---|---|---|---|---|---|
Player | 5 star plays/ opps/ pct | 4 star plays/ opps/ pct | 3 star plays/ opps/ pct | 2 star plays/ opps/ pct | 1 star plays/ opps/ pct |
Upton | 1/ 8/ 12.5% | 4/ 9/ 44.4 | 4/ 6/ 66.7 | 12/16/ 75.0 | 15/15/ 100.0 |
JD Martinez | 0/ 8/ 0.0 | 0/ 3/ 0.0 | 3/ 5/ 60.0 | 3/ 4/ 75.0 | 7/ 7/ 100.0 |
Collins | 0/ 7/ 0.0 | 0/ 3/ 0.0 | 2/ 2/ 100.0 | 4/ 7/ 57.1 | 17/18/ 94.4 |
Jacoby Jones | 0/ 6/ 0.0 | 1/ 1/ 100.0 | 2/ 2/ 100.0 | 3/ 3/ 100.0 | 6/ 6/ 100.0 |
Mahtook | 0/ 4/ 0.0 | 0/ 1/ 0.0 | 1/ 2/ 50.0 | 3/ 4/ 75.0 | 4/ 5/ 80.0 |
Aducci | 0/ 6/ 0.0 | 0/ 1/ 0.0 | 1/ 1/ 100 | 0/ 1/ 0.0 | 5/ 5/ 100.0 |
Romine | 0/ 5/ 0.0 | 2/ 7/ 28.6 | 0/ 0/ NA | 3/ 3/ 100.0 | 5/ 6/ 83.3 |
What we see here is that Jacoby Jones has made every play on balls hit where outfielders make the play at least 20 percent of the time. Justin Upton has also excelled, while others are closer to average.
Another chart shown at Baseballsavant.com shows how often a fly ball is turned into an out based on “hang time.” While a ball that stays in the air five seconds will be an out over 80 percent of the time, a three second fly ball is converted to an out only 23 percent of the time. Hang time combines trajectory, distance and speed of the ball.
Catch probability does not account for difficulty in making plays at the wall (or crashing into a team mate), and it is limited to outfielders so far. It is a big step forward in technology that helps make defensive metrics more accurate.