Remember when it was the end of spring training, and the Orioles were exploring ways to not have Hyun Soo Kim be on the roster? My computer tells me it’s June 6, and I have reason to believe my computer, and if you set a low enough minimum, Kim owns the highest wRC+ in Baltimore’s lineup. He ranks 10th in all of baseball, and though that 10th sandwiches him between Tyler Naquin and Steve Pearce — it’s early — it’s not hard to draw parallels between 2016 Kim and 2015 Jung Ho Kang. Kim presumably isn’t this good, but he’s talented, and now he’s showing that he can hold his own against big-league competition. It didn’t look like that in March, but March has lied to us before.

Let’s dig into Kim just a little bit. He’s batted 78 times, and nearly every single one of those plate appearances has come against a righty. So, that’s a factor. And he’s hit a ton of ground balls. He has one home run, but if Kim keeps putting balls on the ground, that limits his power upside, obviously. Now, Kim hasn’t chased out of the zone very much. He’s also been better than average at putting the bat on the ball. And we can address the grounders head-on. With help from Baseball Savant, naturally.

Statcast doesn’t quite record every batted ball, but it gets most of them, and Kim ranks in the 88th percentile in average batted-ball speed. That seems great, but then there’s this: Kim has hit his grounders harder. As a matter of fact, Kim leads baseball in average grounder speed. Leads baseball! Higher than 96 miles per hour. It’s good to lead in a contact metric, but then, hard grounders aren’t necessarily better than soft grounders. Generally speaking, hard contact is nearly wasted on a ground ball.

That’s the downside — Kim hits hard grounders, instead of hard flies. Now here’s the upside of that downside. Kim also has baseball’s fourth-highest average grounder launch angle. That might sound kind of funky, but Kim’s grounders so far have an average launch angle of -3.3 degrees. The league average is -9.9 degrees. So of Kim’s grounders recorded, they’ve been closer to the line between grounders and line drives. Here’s how the league has done, in batting average, by grounder launch angle:

• -5 to 0 degrees: .315 average
• -10 to -6 degrees: .200
• -15 to -11 degrees: .154
• -20 to -16 degrees: .120
• -25 to -21 degrees: .082

The closer you get to a flat exit, the more productive the batted ball. And that’s intuitive, I think, because those are the most like line drives, and defenders have the least time to react. Here’s Kim against Dellin Betances last Friday:

That was recorded as a ground ball. As a bonus, that features Kim making solid contact against elite-level velocity, but the point is that while Kim hit a grounder, he really hit more of a line-drive grounder. And there’s evidence that could be a skill of his. If this were to keep up, Kim wouldn’t hit a bunch of dingers, but he would hit liners and he’d end up with a strong average and BABIP. He’s used a lot of his hard contact on grounders, which is bad, but those grounders have almost been like liners, which is good. You understand. You’re a smart person!

Not every Kim batted ball has been recorded, but on the 15 without Statcast readings, Kim has gone 5-for-15 with two doubles, so I don’t think we’re missing a bunch of horrible contact. And data points get dropped for every hitter. We can use only what we have, and for Kim, there’s a good thing about the bad thing. And, you know, maybe in time he’ll start to elevate the ball even more. I don’t know what he’s going to do, and last year Kang hit more fly balls after April and May. If Kim puts that contact in the air, that’s great. If he stays as he is, that’s fine. Hyun Soo Kim is looking like he can cut it. Take that, March.

In The Only Rule Is It Has to Work, the excellent new book by Ben Lindbergh and Sam Miller, the authors served as the baseball operations department for the Sonoma Stompers, an independent league team in the Pacific Association of Professional Baseball Clubs. The analytically inclined writers-turned-executives were given freedom to make roster, lineup and strategical decisions based on data, and among the most radical ideas explored in the book is the implementation of a five-man infield against an opposing player named Scott David.

The Stompers deploy a five-man infield against Scott David of the Pittsburg Diamonds, with Mike Jackson Jr. on the mound. (Source)

David is one of the best hitters in the league, one with seemingly no exploitable flaws in his approach, and the authors were struggling to come up with anything substantial to provide their pitchers in a scouting report. That is, until an off-hand remark was made about the possibility of enacting a five-man infield, and the realization that David was, in fact, the ideal candidate for the radical defensive alignment, for four key reasons:

1. He hits a bunch of ground balls
2. He sprays those ground balls all over the infield
3. He is an effective ground ball hitter
4. He has very predictable fly ball tendencies

As soon as I finished the chapter, I knew I needed to find Major League Baseball’s Scott David. Into the numbers I dove.

Using the FanGraphs leaderboards and BaseballSavant, I was able to put a number on each of the four tendencies above. Step one, ground ball rate, is simple enough to find. For step two, I calculated the difference between pull ground ball rate, and opposite-field ground ball rate to serve as a proxy for how often a player sprays his ground balls. For step three, I simply used ground ball OPS — no point in bringing in an outfielder if the player’s ground balls aren’t doing any damage. And for step four, I found the absolute value of a player’s pulled fly ball rate as a way to highlight predictable fly ball tendencies. Then I summed the z-scores of each of the four numbers to come up with a “Five-Man Infield Score.”

In the top five, we find guys like Eric Hosmer and Nori Aoki, but doing this to a lefty, as Lindbergh and Miller did, is admittedly more dangerous due to the exposure of the right field corner for an easy triple. Right-hander David Freese pulls too many of his ground balls; a normal infield shift will do the trick for him. Jean Segura is a decent candidate, though his ground ball rate and spray tendencies are not nearly as extreme as the number one name that pops up on our spreadsheet, far and away the most ideal candidate in Major League Baseball for the five-man infield: Howie Kendrick.

Look no further than his spray chart for convincing. The yellow dots are rough estimates of the optimal positioning against Kendrick in the proposed five-man infield:

Kendrick’s hit a ground ball on 65% of his balls in play, one of the highest rates in baseball. He hits them all over the infield, so shifting him with four infielders is impossible, and while Kendrick has been a pretty poor hitter this year, his OPS on ground balls is actually above-average, so there’s still hits to be taken away here. And as we can plainly see in his spray chart, there simply hasn’t been a need for a left fielder against Kendrick, so bring him in to play behind the second base bag, shift the the center fielder over slightly, and, voila! We’ve got a five-man infield, perfectly designed for Howie Kendrick, the only obvious candidate for a five-man infield in today’s MLB.

Now who’s going to be the first team to do it?

David Laurila published an interview with Lance McCullers last Friday. On Sunday, McCullers made his latest start, and here is about 1% of it:

Here is a view of the same thing happening, only taking place over a greater amount of time:

Good result. Good curveball! Nothing new there — McCullers tends to get good results, mostly because he throws a good curveball. His curveball bears a strong resemblance to that thrown by Craig Kimbrel, the difference being that McCullers is a starting pitcher, which is nuts. He’s not a two-pitch guy, but everyone knows the curveball is his weapon; as a rookie, his curve was worth about 18 runs better than average, while his fastball and change combined to be worth about -7.

You can’t learn a lot from one video clip. It’s always helpful to establish context. So, here’s the whole point of this. Last season, among starters, McCullers threw baseball’s hardest average curveball, by 0.1 miles per hour. In curveball rate, he ranked sixth. Good curveball, hard curveball, used it a lot. Moving on.

This season, among starters, McCullers has thrown baseball’s hardest average curveball, by 2.0 miles per hour. In curveball rate, he ranks first. He ranks first by more than seven percentage points. McCullers has thrown a curveball 49% of the time, and out of his five starts, his lowest rate is 42%. You think Rich Hill curveballs a lot? You think Drew Pomeranz curveballs a lot? They most certainly do, sure, but not like McCullers. No starter throws curves as often as he does, and no starter throws curves as hard as he does.

For what it’s worth, we have pitch-type information stretching back to 2002. The highest curveball rates on record for starters:

1. Lance McCullers, 2016, 48.5%
2. John Stephens, 2002, 46.0%
3. Phil Irwin, 2013, 41.9%
4. Pat Mahomes, 2003, 41.7%
5. Rich Hill, 2015, 41.5%

McCullers also ranks first in velocity, assuming this is a data glitch:

Compared to last year, McCullers has thrown plenty more curves. Statistically, that’s sensible, because the curve is his best pitch. Even now, the curve still has a strongly positive run value, while the other pitches don’t. It’s interesting to observe that the whiff rate at his curveball has only gone up, and rather substantially so. His fastball is like a secondary pitch at this point. Speaking of which — he’s throwing his fastball a little slower. He’s throwing his changeup a little slower. The curveball is harder. Part of this is probably just having a harder curveball, and another part is probably favoring the sharper curve over the loopier curve, as McCullers discussed with Laurila. He says he’s got two varieties of the breaking ball, and evidence suggests he’s been using more of the hard one.

So Lance McCullers is highly atypical. Or, he’s been so, so far. It’s worth noting it hasn’t all been good news — his ERA is over 4, in part because his walk rate has almost doubled. The strikeouts are up, and the grounders are up, but McCullers is searching for a groove. If and when he finds one, maybe it’ll feature fewer curves. Maybe the curves will be slower. I can’t tell you exactly where Lance McCullers is going to settle.

But the version we’ve seen this season? Haven’t seen a starter quite like this. Not, at least, for a long, long time.