They secretly replaced Yankee Stadium … and other lies we can prove using real stats
In the first few weeks after the baseball season begins next month, we will all be on our guard for bad arguments propped up by small samples of data. We might be tempted to fall for On Pace and What’s Changed, but we also know that a few weeks of stats often — usually? — lie. We, being sophisticated scholars of the truth, will wait a few months, thank you very much.
But the greatest trick the devil ever pulled was convincing the world he had a large enough sample size. Baseball statistics can keep lying all year long, if you let them. A whole season of stats, across the entire league, can still tell a lie or two.
What follow are lies. We repeat: They are lies! We are going to tell you things that are not true, using data that technically is. Do not be fooled. We are lying to you. Although, maybe, there’s a bit of truth to take away from each.
The evidence: OK, so: By May 2009 — just one month after the Yankees opened The House That George Built — the new park’s reputation was already established: It “has become one of the biggest jokes in baseball,” according to Peter Gammons. The park’s dimensions were said to exactly replicate those of its predecessor, but analysis of satellite images cast doubt on that claim, and right field in particular seemed shorter than the old park. “I’m tired of people saying it’s too early, we don’t have enough games,” Gammons said at the time. “We have enough games.”
And, indeed, for the next decade the park would consistently play as one of the most hitter-friendly in baseball, and as probably the most homer-friendly park. When Yankees right-handers put the ball in the air to right field, they slugged 200 points higher at home than elsewhere. When Yankees lefties pulled the ball in the air, they slugged almost 300 points higher at home. And the team’s opponents showed similar splits, hitting — over the course of the first decade — 200 more homers against Yankees pitching in New York than elsewhere. Year after year, Yankee Stadium ranked near the top of offensive park factors:
In the first year, the Yankees and their opponents hit 12% more homers in Yankee Stadium than elsewhere, and so on. And then, suddenly, unexpectedly, inexplicably — unless you know about the secret army of construction workers — that collapsed last year.
Yankee Stadium’s run-scoring park factor was the second lowest in baseball, ahead of only Oracle Park in San Francisco. Yankees hitters hit worse at home than on the road for the first time since Yankee Stadium opened in 2009. Their pitchers, meanwhile, showed their biggest home-field advantage ever, with an ERA at home almost a run and a half lower than on the road. Fifteen Yankees batted at least 50 times at home, and 10 of them hit worse at home than they did on the road — despite having home-field advantage. Poor Didi Gregorius hit .196/.226/.345 at home, .273/.317/.523 on the road. On the road he was a star. But at home, stifled by Yankee Stadium’s suddenly harsh environment, his free-agent case shriveled.
The truth: But, wait. Did the Yankees do worse at home than they did on the road, as we kept stating above? Or did they do better on the road than at home? The answer is, of course, yes to both — it’s the same fact, stated differently. But the framing changed the takeaway.
All of the evidence (for this lie) has been based on two numbers in relationship to each other. Yankee Stadium looks like it’s less hitter-friendly because the same hitters hit worse there than they did elsewhere. But the abnormal performances weren’t in Yankee Stadium (with the exception of Gregorius, cherry-picked for just that purpose). The abnormal performances were everywhere else.
On the road, Yankees hitters were transcendent. They hit the second-most road home runs in history, by any team. And, on the road, their pitchers were awful: They allowed the most road home runs by any team in history. Two extreme outliers.
At home, the Yankees were pretty much what we would have expected. They hit a ton of homers — tied for the second most in the majors, and one dinger shy of their franchise record — just like a team with Aaron Judge, Gary Sanchez, and sometimes Giancarlo Stanton should do. Their pitchers, meanwhile, allowed the 14th-most homers in the league at home. In 2018, with most of those same pitchers, they finished … 14th in the league, too.
So, yes, something strange happened to the Yankees last year. But it wasn’t what they did at home. It was what they did on the road. That leaves two possibilities: One is that, by a couple of those strange flukes that happen in baseball in relatively small samples, the Yankees’ hitters wildly outperformed their skill level on the road for a season, while by remarkable coincidence their pitchers wildly underperformed their skill level on the road. This is a little bit hard to believe. But it’s not that hard to believe.
The other possibility is that Yankees hitters really were the historic juggernaut they appeared to be on the road, and Yankees pitchers were the disaster staff they appeared to be on the road — and that these true talent levels were disguised by some strange, out-of-character Yankee Stadium stinginess (or the new dimensions of a secretly constructed new ballpark). This is hard to believe, and stays hard to believe no matter how long you keep thinking about it.
The evidence: Last year, Garcia scored 47% of the time he reached base for the Chicago White Sox, not counting home runs. That was the highest rate in baseball, and — relative to the league average rate (31%), it means Garcia scored around 30 more runs than an average runner would have. Thirty runs above average is about what Nolan Arenado added with his bat last year, and what Ozzie Smith added with his glove in his best season. But the weird thing is, Garcia isn’t that fast — he ranks around 100th in the majors in sprint speed, just behind Hunter Pence and Jose Altuve — and he isn’t very aggressive, either. His 15 stolen bases last year tied him with Shin-Soo Choo and Cody Bellinger for 33rd in the majors.
And yet he was the best scorer in baseball, and one of the best this decade. The lone baserunner to score more frequently in the 2010s was Dee Gordon, in 2017 for the Miami Marlins, at 48% of his appearances on base. Gordon, by contrast, was that fast (17th in sprint speed) and he was that aggressive, stealing 60 bases to lead the majors. He was also batting at the top of the National League’s best offense that season, in front of Christian Yelich, Marcell Ozuna and Stanton during his 59-home run season. It all made sense. Garcia, by contrast, batted at the top of a below-average offense, for a White Sox team that scored only 708 runs, the third-fewest in the American League. Despite this lack of offensive support, he memorized the route from first to home and made it often. He’s the best!
The truth: OK, Garcia really was a pretty good baserunner, overall. His 27 “bases taken” — on wild pitches and passed balls, sacrifice flies and balks — was the third most in baseball last year. He took the extra base on hits — going first to third on a single, for instance — 52% of the time, not among the league leaders but well above average. His baserunning WAR was, at 4.5 runs above average, the ninth highest in baseball.
But 4.5 runs is a far cry from the 30 runs his runs-scored percentage implies. So where are all those coming from? Two main things.
The first is that the batters behind him were, although nothing special on paper, surprisingly special when Garcia was on base. Chicago’s Nos. 2, 3 and 4 hitters were collectively worse than the league average in those spots. But they raked when Garcia (specifically Garcia) got on base:
Jose Abreu: .345/.368/.527 with Garcia on base
Yoan Moncada: .366/.460/.659
Tim Anderson: .342/.378/.487
James McCann: .364/.417/.955
Those are the White Sox who most often batted behind Garcia. The club as a whole hit .322/.377/.535 with Garcia on base. Weighted for their overall season lines, the White Sox would have been expected to hit just .283/.329/.478. Through some unexplained force, Leury Garcia’s presence on the bases turned everyone on the White Sox into Vladimir Guerrero.
The second is that Garcia had a knack for getting on base with nobody out. A runner who reaches first base with none out is about 50% more likely to score than a runner who reaches first base with one out, and about three times more likely than a runner who reaches with two out. Garcia usually batted leadoff, so more of his plate appearances came with nobody out. Furthermore, though, his on-base percentage with nobody out was, at .341, much better than it was with one or two out (.280). Put those two facts together, and 54% of the time he reached base it was with nobody out. Only one other player in baseball was even as high as 50%. In a small and hidden way, his ability to reach with nobody out made him a little bit clutch.
The evidence: From 2012 to 2015, Felix Hernandez was one of the best pitchers in baseball. He went 58-34 with a 2.92 ERA, getting Cy Young votes all four seasons. His average fastball was 92 mph.
Then 92 became 90.5 (in 2016 and 2017), and then 89.5 over the past two years. It has been horrifying to watch: His ERA and FIP rose in each of those years, and last year he won only one game. His career as a Mariner ended, and in January he signed a minor league deal with the Atlanta Braves.
It’s pretty obvious that a fast fastball is better than a slow fastball, and this simple chart showing performance against fastballs at each velocity in 2019 demonstrates it:
With only tiny variations, every tick of velocity above 89 mph sinks offensive performance further. But notice the curious detail on the far left of the graph: Taking a tick away from 89 seemed to make it harder to hit, too. Indeed, in 2019, fastballs from 87.00 to 87.99 mph produced the same offense, basically, as fastballs from 92.00 to 92.99.
In baseball, there’s a term for those slow fastballs: “Below hitting speed.” It’s often used as a backhanded compliment for pitchers, but it’s also often used sincerely to describe pitches that are harder to hit than their slightly faster brethren. Hitting is a nearly impossible task made possible only through the pattern recognition that batters develop over the course of tens of thousands of pitches seen. Because super-slow fastballs are so unfamiliar to major league hitters, they don’t fit into that pattern, and they’re hard to slow down for.
Here, for instance, is Bruce Chen — who went 82-81 as a major leaguer despite a mid-80s fastball — talking about it in 2014: “I’ve had a lot of guys tell me that, to ‘keep throwing below hitting speed.’ I have noted the harder I try to throw, the less favorable results I have. Hitters have their timing mechanisms. That’s why radar guns are so important to them. They want to know how hard, so they can time it against, ‘OK, this is what I usually do against a guy who throws this speed.’ And they’re used to more speed than mine.”
So, the lie goes, 87 is the same as 92. And when Hernandez was at 92, he was a Cy Young candidate. All he needs to do is get back to 92 — or get down to 87. Simple!
The truth: This is, of course, a fantasy. I asked one pitcher who often worked in the high 80s whether he ever felt like he was getting an advantage from throwing that slow. “Nope. Ha.”
There are a few pitchers who succeeded in the majors throwing in the mid-to-high-80s last year: Kyle Hendricks, Mike Leake and Marco Gonzales, the three lowest-velocity pitchers to qualify for the ERA title, were all above-average pitchers, and Hendricks is probably one of the 20 best pitchers in baseball. They’re the bulk of the reason the line dips as it moves to the left in that graph up there. But the pool of pitchers who throw 162 innings in the big leagues each year isn’t randomly selected. Those three did it because, out of the thousands of humans in the world capable of throwing in the mid-80s, they and they alone have the whole package to thrive at those velocities. That might mean deception, movement, command, fastballs that “rise” to get above barrels, and well-tunneled pitches that all look the same at the point a batter has to decide whether to swing. It doesn’t mean the lower velocity itself is the advantage, any more than the occasional short NBA player proves the disadvantage of height.
I prefer “lefty who has gotten the most out of what he has”. Living AT or BELOW hitting speed is the single most dangerous thing you can do on a ball field. TRUST ME. #PrayForTheThumber #SouthPause https://t.co/wEYpRnd1Ma
Those few examples should provide a little hope for every great pitcher who loses a bunch of his velocity. It is possible to succeed at those speeds, and with a new organization, new pitching coaches and new expectations, Hernandez will continue the hard work of finding a plan that works for him at those speeds. As part of the right package, working below hitting speed might even provide the occasional small benefit. It’s incredibly difficult, though, and the most likely outcome is what awaits nearly all pitchers some day: obsolescence. The sample size on that one is enormous.
Thanks to Lucas Apostoleris and Baseball Prospectus for research assistance.