Welcome back to Stats All, Folks, my occasional series that focuses on baseball analytics and sabermetrics and how they’re influencing the sport.
The Orioles’ call-up of center fielder Cedric Mullins this past weekend — and their decision to move longtime center fielder Adam Jones to right — has sparked plenty of discussion about the club’s outfield defense. With that in mind, now’s a good time to review a couple of tools used by MLB’s Statcast technology to evaluate outfield defense: Catch Probability and Outs Above Average.
What is Catch Probability?
Catch Probability is the probability that a fly ball will be caught. It’s based on two key factors: how long the ball is in the air, and how far the fielder has to run.
As any baseball fan knows, fly balls come in all different forms. Some hang in the air forever and could be caught by basically anyone wearing a glove, even you or me. (Well, maybe not me.) Others are blistering drives hit to the gaps with little hang time, which almost nobody could catch. And many are somewhere in between those two.
What Catch Probability boils down to is, for any given fly ball, how often has a ball with that same hang time and distance been caught in the past? To determine this, Catch Probability uses Statcast hang time and distance data since 2015. Hang time is the amount of time that elapses from when the pitcher releases the ball to when the ball is caught (or lands safely). Distance is how far the fielder runs from his starting position to where the ball ultimately comes down.
The longer a fly ball is in the air, and the less distance the fielder has to cover to get there, the easier the play is. For instance, a ball that’s in the air for six seconds or longer will almost always be caught, as will a play that requires the fielder to move 50 feet or less. By contrast, if a fielder has only, say, four seconds to run 80 feet, it’s going to be extremely difficult for him to make the play.
Catch Probability is sorted into five categories based on degree of difficulty: five-star plays (plays that have a Catch Probability of 0-25 percent), four-star plays (26-50 percent); three-star plays (51-75 percent), two-star plays (76-90 percent) and one-star plays (91-95 percent).
What is Outs Above Average?
Outs Above Average is a measure of how well — or how poorly — an outfielder turns fly balls into outs, based on their Catch Probability.
Every time an outfielder catches (or fails to catch) a fly ball, his OAA is credited (or deducted) a few points based on how difficult the play was.
For example, let’s say an outfielder misses a fly ball that had a 48 percent Catch Probability. He would have .48 points deducted from his OAA score. This reflects that he wasn’t able to make a play that 48 percent of outfielders would have.
On the other hand, if he had made the catch, he would have .52 points added to his OAA, because he made a play that 52 percent of outfielders wouldn’t have been able to.
In this way, outfielders get extra credit for making strong defensive plays. An outfielder who catches a fly ball with a Catch Probability of just 2 percent would add .98 points to his OAA. By the same token, missing a routine fly ball is costly for an outfielder. If you drop a fly ball that has a Catch Probability of 95 percent, your OAA would take a .95-point hit.
Statcast tracks every outfield play throughout the season to help calculate a fielder’s cumulative OAA. Currently, the best OAAs in the majors belong to the Cincinnati Reds’ Billy Hamilton and the St. Louis Cardinals’ Harrison Bader at +15 each. Last season, Minnesota Twins’ Gold Glove center fielder Byron Buxton led MLB with +29 OAA.
How does the Orioles’ outfield rate in Outs Above Average?
If you’ve watched the Orioles’ outfield defense this season, it probably won’t surprise you to learn that the club ranks last in the majors — by a lot — in Outs Above Average. As of Friday, the day Mullins was called up to the majors, the Orioles had a -22 OAA for the season, six points worse than the next-worst team, the Philadelphia Phillies. The Orioles’ mark was a whopping 41 points worse than the best team in the majors, the Milwaukee Brewers, who had a +19 OAA.
The Orioles have done OK at making the most routine catches (the one-star category of plays), and they rank in the middle of the pack at making the most difficult plays (the five-star category). For four-star plays, the fairly difficult ones (which have a 26-50 percent Catch Probability), the Orioles rank fifth to last in MLB at 22.2 percent.
What has really hurt the Orioles is their inability to making the two-star and three-star plays — the fairly easy and moderate catches.
Two-star plays are those that are supposed to be just a step below routine, with a Catch Probability of 76-90 percent. The Orioles have made only 76.8 percent of those catches, worst in the majors, and are the only team below 80 percent in that category.
Then there are the three-star plays. Those are plays that aren’t exactly easy, but are still caught by the average outfielder at least half the time (51-75 percent). Well, the Orioles are the exception. They’ve turned just 41.3 percent of those fly balls into outs, a horrendous mark. Again, they rank last in the majors in that category, and they’re the only team that’s made less than 50 percent of those plays.
The two Oriole outfielders with at least 100 fielding chances ranked as the fourth- and fifth-worst in the majors in OAA. Adam Jones struggled in center field, amassing -10 OAA in center before shifting to right. Trey Mancini was even worse in left field with a -11 mark.
In center, Jones made all the one-star plays and most of the two-star plays, but his defense dropped off the table after that. He made only 25 percent of the three-star plays (three out of 12) and didn’t make a single four-star or five-star catch in 33 opportunities. It’s an indication that the 33-year-old Jones might have lost some speed and lacks the range he once had, and it supports the idea that shifting to a corner outfield spot was the right move for him and the Orioles.
Mancini has had sort of the opposite problem. He’s actually managed to make a few impressive catches this year — one play in the five-star category, three in the four-star category and seven in the three-star category — but has bungled the more routine ones. He has made only 50 percent of two-star plays (five of 10) and has even botched three of his one-star plays, converting just 82.4 percent. That may be a result of Mancini, a converted first baseman, taking poor routes to fly balls that veteran outfielders would track down with ease.
The struggles of Orioles’ outfielders are nothing new. Last year, Jones had a -7 OAA and Mancini -5, and the Orioles as a group were at -15. In 2016, the Orioles’ OAA mark was -18.
That porous outfield defense is one of the many things that’s gone wrong for the 2018 club. Think of the extra runners the Orioles’ outfielders are allowing to reach base on fly balls that most other teams would catch. It plays no small part in the MLB-worst 639 runs the Orioles had allowed as of Sunday.
How can the Orioles improve their OAA?
The Orioles’ promotion of Mullins might go a long way toward fixing their defensive issues in the outfield. Mullins has a strong defensive reputation and should instantly improve the Orioles’ center field defense, catching up to fly balls that Jones no longer could. And Jones, too, should see an uptick in his defensive numbers by shifting to the less demanding right field. As for Mancini, well, that’s a problem still in need of a solution.
If Mullins plays the type of center field he’s capable of, expect to see the Orioles turn a lot more fly balls into outs for the rest of this season and beyond. And with those potential baserunners wiped off the board, that, in turn, should lessen the pressure on the Orioles’ pitching staff — and maybe lead to an extra couple of wins that the club otherwise wouldn’t have gotten.
It’s not going to salvage the 2018 season, of course. But for an Orioles franchise that once prided itself on playing excellent defense, the addition of Mullins is a big step in the right direction.
After the Orioles’ elimination from the postseason, executive vice president/general manager Mike Elias promised a…
The surprising news that the Orioles were bringing the left-field wall closer to home plate…
After three seasons of a heightened and lengthened left-field wall, the Orioles have decided to…
Question: Rich, any chance Trey Mancini gets a minor league contract and a chance to…
The Orioles' coaching staff for 2025 is close to being finalized. According to an industry…
Question: With all the interest and hype surrounding signing either players already on the team…
View Comments
I don't guess these stats are kept track of at the minor league level? It'll be interesting to see where Mullins stands at the end of the year.
I'd be curious to know how this all translates into defensive runs saved in the outfield. It would be interesting to look at various scenarios where different outfield configurations are in place to look at runs saved verses runs lost to decrease in offensive output. I wonder, for example, if we've cost ourselves games with Mark Trumbo in RF to boost his offensive output because he is less productive as a DH. Games that we might have won with better outfield defense tracking down flyballs and just dealing with reduced productivity from Trumbo as the DH. Remember, for example, on opening day, Gentry made a game-saving catch at the wall in LF that allowed Jones to hit the walk-off HR in the 11th to beat Minnesota. What would our season look like (not that much better, I'm sure) if we had more speed in the outfield but sacrificed some offensive output? How many of the 15-18 games (not exactly sure of the figure, but it's in this range) that were lost by 1 run would have been wins except for a gapper that plated a couple of runs? I didn't look at 2-run games because, frankly, as I looked back at the schedule I realized that we had a few too many blowouts for 1 and 2-run games to have really turned the season around. Still, it would be interesting to try to quantify the games lost due to slow outfield defense.
That would be an interesting thing to quantify, and with so many analytics-heavy teams out there, I'm sure somebody is already attempting to do just that. It's above my pay grade, though.
These kind of stats are one of my problems with over reliance on analytics. The eyeball test can tell you who is a good or bad outfielder. But dividing all chances into five levels of difficulty is obviously at least somewhat subjective. But since the final results are mathematical they MUST be true.
These sorts of stats are less subjective than the eye test. If you want a player to be better, you are going to be biased. If you don't like the guy for some reason, you are also going to be biased. I've read too much about how Adam Jones is still a great centerfielder, and there's only one reason for them to think that...they like the guy.
As for objectivity overall...fans are super subjective...see fans who believe their particular legendary player is "the greatest of all time" or fans who get the fussies when a national sportswriter predicts their mediocre team will be mediocre. These stats are not fool-proof but are better than emotion.
The eyeball test has its limits, though. For every fly ball that's hit, are you watching the outfielders the whole way? Can you tell how far they ran or what route they took? A lot of times, all we see is the end result of the play, and that doesn't come close to telling us the whole story.
Plus, how many teams do we watch every single day besides the Orioles? I don't think I could tell you who all the best and worst defensive outfielders are on the other 29 teams just based on seeing them for a few series a year. That's why defensive stats like this are very helpful. And they're much more objective than the eye test, as JCO noted.
The O's out-performed their Pecota ratings every year under Buck until 2017. So is Buck a much better than average manager (which is being denied now) or are analytics not perfect?
I never said analytics aren't useful. But they are not perfect. Many former players ( and i'm sure former coaches and managers) are questiioning an over-reliance on analytics. Whether it is the way pitchers are used or launch angle.
Well, if you don't like analytics, this isn't good news: https://www.mlb.com/news/major-league-clubs-eyeing-future-of-analytics/c-220934644
Much of it is how they use it. Current players are buying into it, as that story showed.
And any older Oriole fan (like myself) who pines for the good old days, let's not forget who the man who is considered by some to be the father of baseball analytics because he used them extensively before anyone else: Earl Weaver.
And I'll see your former players grousing about it and raise you this: https://www.bloomberg.com/news/articles/2014-08-28/extreme-moneyball-houston-astros-jeff-luhnow-lets-data-reign
Great explanation and I hope you will follow up after we have seen Mullins play for a month. And if DJ Stewart comes up in September?
Are these the kind of analytics Duquette was referring to when laying out the O’s future plans?