The 2019 Yankees are great, but they’re also lucky

Remember how the Yankees won 100 games last year and didn’t sniff a division race? It was frustrating; the Yankees could sweep an August series and look up at the standings, and the Red Sox were still eight games up. It was almost impossible to imagine coming back to take the AL East after May, and sure enough the Yankees, ended up settling for the Wild Card.

One year later. I can only imagine the fans of the Tampa Bay Rays are in a similar position. The Rays are quite good, projected to win 92 games and earn a playoff berth, but every time they do something well, they’re still stuck nine games behind the division-leading Yankees.

We all felt the 2018 Sox had a little bit of luck on their side, but the truth be told, so do the Yankees. At time of writing, they’re 85-47, with a 152 run differential and all but a lock for their first division title since 2012.

When you pull apart the Yankees’ actual performance, however, they’re not nearly as invulnerable as a playoff odds graph might indicate. Their pythagorean record is 80-52, meaning they’ve won five more games than their pure run differential suggests they should have. The Rays, meanwhile, have a 76-56 pythagorean record, exactly what their performance matches. The Yankees have over-performed while the Rays have matched output, and that’s the difference between a nine-game lead and a four-game lead.

If you go deeper into FanGraphs’ BaseRuns records, the Yankees are 10 (!) games better than their underlying metrics indicate. The Rays, by BaseRuns, would have a five-game lead in the AL East, reinforcing the gratitude among Yankee fans that the games on the field are what count in the standings.

Why is this, though? What are the Yankees doing differently than the Rays, that they’re able to outperform these underlying numbers and Tampa can’t?

The bullpen proves a good place to start. A team with a good bullpen will usually over-perform their metrics, since a shutdown relief corps helps preserve one and two-run leads that present the most uncertainty in models like BaseRuns.

The Yankees have had the most valuable bullpen on a per-inning basis in all of baseball. More importantly, perhaps, the Big Four at the back of the bullpen—Aroldis Chapman, Adam Ottavino, Tommy Kahnle and Zack Britton—have combined for a ludicrous 20% strikeout-to-walk rate, the best indicator of a pitcher’s success. Only five starters in MLB have a K-BB% higher than the Big Four’s combined numbers.

When the Yankees have a slim lead, the high-leverage guys in the bullpen are more likely than any in baseball to hold that lead. The other thing that tends to drive over-performance is clutch hitting, and the Yankees lead all of baseball with a .300 average with RISP. They have a 128 wRC+ in those situations, which also tops in MLB.

Hitting with RISP, however, isn’t exactly a skill. It’s incredibly volatile year to year and subject to a whole host of factors. If you take a look at the Yankees’ clutch score, a metric FanGraphs produces that measures a performance in high-leverage plate appearances vs low ones, the Yankees have been dramatically up and down over the last decade:

The last time the Yankees were the best team in baseball and won the World Series, they were an “unclutch” team; the last time they were as “clutch” as they are this year was 2013, not a year fans tend to hold in the highest regard.

Even across all of baseball, hitting in clutch situations is volatile, which is why it’s so hard to pin it down as a repeatable, observable skill. If hitting in clutch spots was a skill, we’d see teams that return virtually identical lineups—like the Red Sox—repeat their clutch performance year over year. Instead, this year Boston falls 17 points worse with RISP than they were last year, the season they ran over all of baseball en route to the World Series.

So part of the Yankees’ over-performance is due to skill. The bullpen performance is much more tangible, and indeed a stated goal of this Yankees administration every offseason. The other part is far more volatile, and you can call it luck or positive random variations from the mean, or whatever you want. It’s yet to be seen how sustainable that variation can and will go, but we saw last year that the Red Sox could keep a string of luck all the way to the World Series.