The first time he did it, we thought, well, even the weakest Major League defender is still a Major League defender. The second time he did it, we thought maybe he was just making an easy play look harder than it needed to be, as poor defenders often do. But then came the third time, in the biggest spot, and all of a sudden, we¡¯re typing words we never thought were possible: Nick Castellanos is playing like a defensive superstar this postseason.

It would be an incredible month for anyone, but it¡¯s simply astounding coming from him, because it¡¯s neither inaccurate nor unkind to point out that Castellanos¡¯ defensive deficiencies have followed him around since the moment he made it to the big leagues. From 2014-17, with the Tigers, he was rated as baseball¡¯s weakest regular defensive third baseman, which was enough to get him moved to the outfield full time. He's since rated as baseball¡¯s weakest regular defensive outfielder from 2018-22.

This year was no outlier for Castellanos. Pressed into more regular outfield duty than anyone expected -- when he signed, he was expected to spend much of his time at designated hitter, before Bryce Harper¡¯s elbow injury forced him into the bat-only role and Castellanos back to the outfield -- he finished as the second-weakest right fielder in the game, per Statcast¡¯s Outs Above Average. (That he hasn¡¯t been charged with an error in more than a year speaks far more to the limits of both his range and the error stat itself than it does his sure-handed fielding.)

It¡¯s not that Castellanos is too slow to play the outfield, as his top speed is better than about 60% of baseball, and similar enough to Aaron Judge, who competently handled center this year. It¡¯s mostly about poor reactions, where his outfield jumps -- or feet covered in the first three seconds, in the right direction -- are among the weakest in the game. That¡¯s where speed comes in, because you can outrun bad jumps with great speed, as his Phillies teammate Matt Vierling does. But the combination of poor jumps and non-elite speed makes for difficulty getting to balls you need to get to.

All of which was the narrative of the month, when it was noted that while the Phillies had indeed done a solid job of improving their defense up the middle, the four corners remained an extreme weakness. That assumption hasn¡¯t exactly been erased in October, of course, but when you see Castellanos make catch after catch after catch, you want to know how, why and whether they were as impressive as he made them look -- or if he was just making them look that way with slow reactions.

By the Statcast metrics -- which look at distance, time and direction, and display how often similar opportunities are turned into outs by other fielders -- that¡¯s not what happened. Castellanos is making strong plays on difficult chances. In the regular season, Castellanos had 297 opportunities to make a play, which he did only 81% of the time. Include the postseason and that¡¯s over 300. By the Statcast numbers, his four best plays of 2022 have all come since the playoffs began.

If there¡¯s a thing as ¡°clutch fielding,¡± anyway, this is it. Let's dig deep on the three most prominent plays.

Great catch No. 1: Oct. 11, NLDS Game 1

William Contreras batting, bottom of the 9th, bases empty, Phillies up 7-6
55% Catch Probability

The situation:

After dispatching the Cardinals in the Wild Card Series, the Phillies traveled to Atlanta for the best-of-five Divisional Series, and Game 1 wasn¡¯t going well. Staked to an early 6-1 lead, Ranger Su¨¢rez couldn¡¯t get out of the fourth, and Matt Olson¡¯s three-run home run pulled Atlanta to within a run with one out in the ninth. Contreras stepped up as the tying run, with Travis d¡¯Arnaud, who had homered earlier in the game, on deck.

Zach Eflin got ahead 0-2, then delivered a cutter below the zone, which Contreras sliced to right field. Castellanos, playing 298 feet deep -- a few steps deeper than the usual 293 feet that right fielders positioned themselves against Contreras in Atlanta this year -- was presented with the following opportunity: 64 feet to cover, 4.1 seconds to do it. Similar opportunities are turned into outs 55% of the time.

If that doesn¡¯t sound terribly impressive -- a 50/50 chance, more or less -- realize that Castellanos had only once previously in his outfield career turned a harder chance into an out. (By comparison, San Diego Gold Glover Trent Grisham, one of baseball¡¯s best defenders, turned 16 such chances into outs this year alone.)

If Castellanos doesn¡¯t catch it, the winning run comes to the plate with one out. (Or worse: If he dives and it gets past him, the tying run either advances to scoring position or perhaps even scores, in the worst-case scenario.) By making the catch, Eflin was left with no one on and two outs, soon ending the game.

What the numbers said:

Castellanos got a decent jump (one foot better than average in the right direction, three feet better than he usually does) and good-enough speed (27.2 ft/sec, almost exactly average) but mostly he ran an extremely direct route, traveling 65 feet to get to a landing spot he¡¯d started 64 feet from.

What Castellanos said:

¡°It's kind of just a fresh start, a clean slate, so to speak,¡± he said. ¡°And obviously these games are really intense. For me, that helps me lock in and kind of slow things down. It's just a lot of fun. Baseball is really, really fun right now.¡±

Well, yes:

That outlook is going to end up being a really interesting part of this. Hold that thought for a minute.

Great catch No. 2: Oct. 28, World Series Game 1

Jeremy Pe?a batting, bottom of the 9th, runner on second, tied 5-5
65% Catch Probability

The situation:

And you thought the Contreras play was a big spot, didn¡¯t you? This time, the Phillies had clawed back from a 5-0 hole against Justin Verlander, of all pitchers, to tie the game 5-5 in the fifth. There it remained until the ninth, when Chas McCormick and Christian V¨¢zquez each struck out against Seranthony Dom¨ªnguez. With two outs, Jos¨¦ Altuve singled and then stole second base, placing the winning run in scoring position for Pe?a, who has been one of the better Houston bats in the postseason.

On a 1-1 count, Pe?a lofted a soft fly ball to right field. It wasn¡¯t hit hard (68 mph) or all that far (200 feet). It wasn¡¯t notable in any way, except that it would have won the first game of the World Series. It would have, anyway, if not for Castellanos, who slid to make the catch and extend the game. In the 10th, J.T. Realmuto hit the go-ahead home run.

But more notable than the catch itself is where Castellanos even was. The average right fielder this year stood 289 feet deep against Pe?a in Houston, and that¡¯s exactly where Castellanos was on the first pitch: 289 feet. But after the steal, Castellanos moved in by more than 20 feet, so for the 1-1 blooper, he was just 267 feet away.

Had he been 90 feet with the 4.5 seconds he had, the catch probability would be something like 1%, or, let¡¯s just be honest, ¡°impossible.¡± This play was made before the pitch was even thrown.

What the numbers said:

Almost exactly like the play in Atlanta, Castellanos got an exactly average jump and reached an exactly average speed, but he ran a great route, traveling 76 feet to get to a landing spot he¡¯d started 74 feet from. This one, of course, was about the pre-play positioning more than anything.

What Castellanos said:

¡°I don't know, that was just what my instincts told me to do,¡± he said. ¡°I just thought he had a better chance of trying to bloop something in there than torching something over my head. So that was kind of my thought process there, just thought of it on the fly.¡±

Interesting enough to us that Castellanos was shallow more because he thought the ball would go there and less because it would give him a better chance to hold Altuve at third on a single, but in the end, both goals worked together.

Great catch No. 3: Nov. 1, World Series Game 3

Jose Altuve batting, top of the 1st, bases empty, tied 0-0
55% Catch Probability

The situation:

This one almost got lost in the barrage of home runs the Phillies laid on Lance McCullers Jr., but coming as it did on the very first pitch of the very first World Series game Philadelphia had seen in more than a decade, it certainly set the tone for a successful night. It¡¯s almost a carbon copy of the others, in that it was softly hit and Castellanos had to come in. (Which is important, because he¡¯s much better coming in than he is going back.)

What the numbers said:

The speed was unimpressive (25.5 ft/sec) and the jump was slightly below average, but the route was perfect. Perfect. He needed to go 52 feet. He went 52 feet.

So: What changed?

¡°I think that a lot of times I have trouble keeping attention during the regular season every day for nine innings,¡± Castellanos told the FOX broadcast during Game 1. ¡°But with the postseason, this kind of baseball is incredible, you don¡¯t have a choice but to be locked in, watching swings, watching balls come off the bat, I think that¡¯s just kind of why I¡¯m playing better.¡±

"I felt like I read the swing pretty well," he later said, "and as soon as I saw the direction of the ball, I got a good jump on it."

Locked in. Really fun right now. Keeping attention. Got a good jump.

By Castellanos' own words, and the numbers, the difference appears to be about rising to the moment, specifically in terms of getting moving as quickly as possible. He¡¯s not really running faster. He¡¯s either getting better jumps or positioning himself better.

For a fielder without elite speed, it matters. When he gets a good jump on a ball, it looks like this, when he got off to a start four feet better than average, putting him in position to rob Starling Marte in May. When he doesn¡¯t, like on this Matt Olson double (19 feet worse than average) or this Nathaniel Lowe single (nine feet worse than average), he¡¯s got no shot on otherwise catchable balls.

When we talk about ¡°clutch¡± performances in the postseason, we almost always talk about hitting, or pitching. Maybe -- maybe -- in rare instances, it¡¯s about baserunning. Almost always, it¡¯s a question about whether "clutch" even exists, or if it¡¯s just great players doing great things at the right time. (Does anyone plan to explain to the great Justin Verlander that he may, in fact, not be clutch? Did Bryce Harper only learn to be clutch after signing with Philadelphia?)

Maybe, this year, we¡¯re learning what that means. Castellanos, a much-maligned fielder, is rising to the occasion. Not with the bat, not really; he¡¯s hitting just .197/.246/.262 this postseason, without a home run, which sort of raises the question of when and what 'clutch' is and how it can be applied. It is, improbably, value added with the glove.