What Was Marcell Ozuna Thinking?
With two outs in the bottom of the 10th inning on Monday, the Cardinals found themselves up by one run. Matt Bowman, the Cardinals pitcher, put himself in a little bit of trouble when Rhys Hoskins hit a single and then advanced to second on a groundout. Cardinals manager Mike Matheny put Bowman in considerably more trouble with the dubious decision to walk Carlos Santana and put the winning run on base in order to try for the double play. Bowman did not get the double play, instead striking out Jesmuel Valentin. That brings us to Aaron Altherr, the game’s final batter.
The win-expectancy chart provides a pretty good idea of what happened on that play.
Source: FanGraphs
If the graph doesn’t help enough, here’s a small clip of what transpired.
Marcell Ozuna dove for the ball and, by missing the catch, allowed Hoskins and Santana to score and win the game for the Phillies. Mike Matheny defended the aggressive play, because that’s what a manager is supposed to do. That doesn’t prevent us from asking the question, though: just how badly did Ozuna screw up by trying to dive for a catch he wouldn’t end up making?
A closer look at the play shows that the ball landed a bit short of Ozuna’s glove, and he would have needed an extra step to make the play.
To add injury to insult — or the reverse, depending on how you look at it — the ball caromed off Ozuna’s face and past him, which made the possibility of recovering even more distant for Santana. According to Statcast, Ozuna had almost no shot at the play.
Here's what disaster looks like if you're the Cardinals.
Marcell Ozuna aggressively goes for a ball with a 4% catch probability, and it gets by him for a walk-off two-run double. pic.twitter.com/Dwdda1WkB7
— Andrew Simon (@AndrewSimonMLB) June 19, 2018
That 4% figure is based on the average outfielder. Given that Ozuna possesses above-average speed, though, and is generally regarded as a good defender — combined with the fact that Ozuna got pretty close — the probability of Ozuna successfully making the grab were probably a little greater than 4%, though not too much greater. To assess Ozuna’s decision on the play, we have to account for the consequences of missing the catch plus three other variables: the chances of Ozuna making the catch, the chances of Ozuna getting in front of the ball were he to have taken a more conservative approach, and the chances of the ball getting by Ozuna like it did.
First, lets deal with the alternate universe where Ozuna makes the safer play and just tries to get in front of the ball. For the sake of argument and the example below, let’s say that there’s a 20% chance that the ball gets by Ozuna no matter what he does and that, in each of those cases, Santana scores the winning run. In that scenario, the Phillies’ win probability goes up to 100%. In this example, Ozuna gets in front of the ball 80% of the time with Santana stopping at third and Altherr at first. In that scenario, the Phillies’ win expectancy is 68.5%. That means, the safe play scenario looks like this:
| Chance of Play | PHI Win % | WP Value | |
|---|---|---|---|
| Keeps Ball in Front | 80.0% | 68.5% | 54.8% |
| Ball Gets By | 20.0% | 100.0% | 20.0% |
| TOTAL | 74.8% |
We will use that as our baseline to judge Ozuna’s decision. We can now compare it to a scenario where Ozuna dives for the ball. Let’s say he has a 10% chance of catching the ball, and let’s say that, in the other 90% of the plays, the ball gets by him two-thirds of the time, while one-third of the time it hits his body or hat or face and bounces in front of him to prevent Santana from scoring. In this case, making the catch would be worth 74.8% WPA, as the Cardinals would then win. If he stops the ball, that is worth 6.3% WPA, as that is the difference between the baseline and runners on first and third. If the ball gets by him, that’s worth -25.2% WPA, because the Phillies just moved the baseline of 74.8% to 100%. Here’s what that calculation looks like.
| Outcome Chance | WPA | WP Value | |
|---|---|---|---|
| Makes Catch | 10% | 74.8% | 7.48% |
| Keeps Ball in Front | 30% | 6.3% | 1.89% |
| Ball Gets By | 60% | -25.2% | -15.12% |
| TOTAL | -5.75% |
Given the above parameters, the Ozuna decision is one that costs about 6% of WPA. By that estimate, he should have made the safe play. For comparison’s sake, let’s say the play was a bit of a fluke and that, in two-thirds of the plays where Ozuna misses the catch, the ball stays in front of him, while the ball gets past him and ends the game in only one-third of the diving plays. This is what that looks like.
| Outcome Chance | WPA | WP Value | |
|---|---|---|---|
| Makes Catch | 10% | 74.8% | 7.48% |
| Keeps Ball in Front | 60% | 6.3% | 3.78% |
| Ball Gets By | 30% | -25.2% | -7.56% |
| TOTAL | 3.70% |
If you believe this scenario to be true, then Ozuna did in fact make the right play. Whether you believe Ozuna played the ball correctly depends on your assessment of the variables involved. If you believe Ozuna had a 15% chance of making the catch, but there is only a 10% chance of losing on the safer choice and the odds of the ball bouncing away from Ozuna were 50/50 if he missed, then Ozuna made the right call at basically the break-even decision point.
The table below details a whole host of probabilities and the break-even number for the ball needing to stay in front of Ozuna for the play to be a good choice.
| Ball Gets By on Safe Play | Catch Chance | Ball Needs to Stay in Front At Least |
|---|---|---|
| 10% | 10% | 66% |
| 10% | 15% | 51% |
| 10% | 20% | 34% |
| 20% | 10% | 52% |
| 20% | 15% | 39% |
| 20% | 20% | 21% |
| 30% | 10% | 40% |
| 30% | 15% | 27% |
| 30% | 20% | 9% |
Keep in mind, some of these scenarios just aren’t plausible. If the ball is going to get by 30% of the time if Ozuna plays it safe, he can’t possible believe he is going to keep the ball in front of him 40% of the time if he dives for the catch. That middle scenario seems sort of reasonable, with a 20% chance that the ball gets by if he plays it safe, a 15% chance of a catch if he dives, and basically a 40/60 split of the ball getting past him if he misses.
Statcast’s 4% catch rate makes it look like a terrible decision, but Ozuna doesn’t know it is a 4% possibility when he breaks for the ball. He has to make the call on the fly. Keep in mind that Altherr’s ball traveled only 272 feet, but similarly hit balls by launch angle (between 17 degrees and 19 degrees) and exit velocity (between 94 mph and 95 mph) generally travel around 25 feet farther than the ball for which Ozuna dove. It’s reasonable to think Ozuna thought the ball might travel just a few feet farther with a tiny bit more hang time. That 4% catch rate is based on needing to travel 47 feet in 3.1 seconds. If Ozuna had needed to travel 45 feet and had 3.2 seconds, the catch rate jumps up into the 20s and makes diving for the play an easy call. Not making the catch ultimately cost the Cardinals the game, but based on the information Ozuna had at the time, it isn’t clear he made the wrong choice.
Craig Edwards can be found on twitter @craigjedwards.
I really liked this analysis. I severely doubt anything like the words and numbers above went through Ozuna’s mind but that he can pretty much intuit most of this on the fly and arrive at a coin flip decision is why he is a world class outfielder. The only really wrong decision here would have been indecision via getting handcuffed and caught between playing this safely and going for the game winning catch while realizing neither the risk aversion of the former nor any chance at the latter.
LOVE THIS!
The development of all the new statistics supports this. No stat, new or old, is the be-all-end-all stat that describes X perfectly. Utilizing them and supplementing them with evidence beyond their inputs increases their value and usefulness.
And it pushes the development of them – one day, we may see sprint speed in catch probability.
Ozuna had about 2 seconds to process all of this and make his decision. A general rule of thumb in extra innings is that the visiting team goes for the win and the home team goes for the tie.
Given the Cardinals’ slim chances of winning throughout the game, and the fact that they had already used 2 of their best 3 relievers and the other one was likely unavailable, I think their chances of winning had he played it safe and tried to prevent the second run from scoring was very slim.
I have to admit I kind of skimmed the probability chart, but I still love this type of analysis–or at least I love that it is possible and that I can find it on Fangraphs.
Altherr is a righty so it’s very possible if not likely that the ball was hooking and Ozuna’s brain couldn’t calculate that fast enough and thought he could make the catch. It is sooooooo hard to not try to catch a ball when you spend 99.9% of the time trying to make the catch.
Not only was that ball hooking, it looks like it’has some nasty topspin. Just look at the path of the ball right after it hits its apex. That thing dove like Kluber curveball. A sliced ball has some side/backspin to it, a hooked usually has side/topspin on it. This is something you learn by experience, but also coaches tell you this starting in high school.
The closer the ball is to the foulline, the less true the ball is travelling. Yes, he only had milliseconds to decide what to do, but this should definitely have bee in the back of his mind while going for that one..
I believe Ozuna said the ball “died” on him supporting his view that he had a better than 4% chance
The reason why Ozuna even got on stage last night was the the again questionable bull pen switches Mr. Philly Mgr. Screwed up again. Oh, look again tonight ( Tuesday)and he now has a streak going of poor bull pen decisoning. He is lost in his own dugout.
It’s farther, not further, the ball flies farther, not further. There’s nothing wrong with what Ozuna did. He tried to make an aggressive play on a ball. He’s a baseball player. His own manager praised it (someone who’s been in the game his entire life). There’s no amount of analysis that can fault Ozuna for this play. What was Ozuna thinking? He was thinking of catching that baseball, period.
Welp, there goes the Pulitzer.
Agreed…This was a lot of wasted analysis.
in what world does a ball get by a world class OF playing it safe 20% of the time? that seems like a bonkers assumption. i’d put it at 5% or even less.
it was probably a bad decision.
I wonder to what degree Ozuna’s throwing arm (or lack thereof) played into the decision. Hoskins is hardly a burner, but with Ozuna’s noodle arm, he probably scores from second no matter what if the ball drops. Perhaps Ozuna thought his only chance of keeping the lead was to make the catch. Either way, Matheny deserves more blame for walking Santana, but St. Louis media just wants to pillory Ozuna.
A better question would be “What was Matheny thinking?” Anyone paying attention could see Bowman was not well. His pitchers were unintentionally up in the zone. He was fidgeting profusely with his hands and fingers. The surprise strike out high in the zone was an accident and he never should have been left in to pitch to the next batter. The real problem is that even if Ozuna got that ball on a bounce and held Santana to 3rd base, Matheny would have left Bowman in there to pitch to the next batter with equally disastrous results. Now Bowman is back on the DL. No surprise there to anyone but Matheny.
I don’t think the numbers work out; you need to use differential values.
For the scenario where Marcell plays it safe and turns a 1-run lead with runners 1st/2nd & two outs to a tie game with runners on the corners & two out, 80% of the time the Win Expectancy for Philadelphia improves from 17.4% to 68.5%, or a factor of 40.88. For the 20% of balls that get past him and the home team walks off, WE goes from 17.4% to 100%, or 16.52. That means playing it safe causes a net Cardinal WPA of -57.4%.
Looking at Marcell Dives, catching the ball 10% of the time drops the Phillies WE from 17.4% to 0%, or -1.74. A 30% chance of keeping the ball in front yields a factor of 15.33 and the 60% for the miss gives us a factor of 49.56, for a total Cardinal WPA of -63.15%.
The safer dive has the three situations come out to -1.74, 30.66, & 24.78, respectively, adding up to a Cardinal WPA of -53.7%. Based off of this, it probably makes more sense to make a controlled dive.
If you assume that a reduced chance of making a catch means that there is a greater chance of a complete miss, the break even point for the catch is between 6% & 7%. If the miss percentage remains steady and instead a knockdown is more likely, then it’s between 4% & 5%.
Another way of looking at it is to say that a controlled dive by Marcell gives a 10% chance of a catch & a 60% chance of knocking it down, whereas an aggressive dive would basically guarantee a 60% chance of missing it completely. At what percentage would it be a better call to make an aggressive dive? Somewhere around 24% – maybe Harrison Bader pulls that off.
Marcell Ozuna is a fairly solid outfielder and if he has control of himself, it’s a better percentage play to dive in that situation. If he is going to completely go for it like he did, it’s a horrible idea.
So….what do you gander that Ozuna was thinking on this one?
https://www.mlb.com/brewers/video/aguilars-odd-2-run-double/c-2180064083?tid=8879232