Revisiting Steven Kwan’s Hidden Power

A little more than a year ago I wondered if Steven Kwan might possess a hidden power. One that went beyond the boxscore and the advanced metrics to answer this question: “How did Steven Kwan post a league-average slugging percentage despite the lowest hard contact rate in all of baseball?”

I’d like to revisit that. But first:

A Note!

When suggesting that “...[C]urrent sabermetric models have a gap that cannot be easily filled,” I should have taken more time to say that the gap is so, so small. I meant to offer some work at the edge of established ideas that we tend to agree with. How can we sharpen that edge a little bit? The nice thing about working here is that if it doesn’t cut like we expected we can hone it back to how it was.

SaladCzar also provided some excellent commentary on the piece, noting the following:

There’s a lot of shaky analysis here and an overreliance on a sample size of exactly one.

First is the idea that his overperformance of xSLG is based solely on baserunning is pretty easily checked. Kwan had 225 total bases over 563 ABs for his .400 SLG. An xSLG of .341 would suggest his xTB would be 192, or 33 fewer. Kwan had 32 2B + 3B this year. If he had taken an extra base in every single one of his doubles and triples, a completely ludicrous proposition, we’re still actually one short.

Next is the idea that “Burst” correlates to significantly outperforming xSLG. The top 10 guys in Burst outperformed their xSLG by an average of about 27 points, the top 25 by about 18 points, and the top 50 by about 10 points. For those 50 guys, I’m getting a correlation of .25 between Burst and xSLG - SLG. Being fast seems to lead to outperforming xSLG, but by nowhere near as much as Kwan did.

Kwan is a smart, disciplined hitter. It takes quite a bit of other talents to go from being in the first percentile in hard hit and barrel rate, and third in exit velocity to finish in the 46th percentile in xwOBA. And yes, he absolutely is going to get a few, emphasis on few, extra bases, helping him potentially outperform his xSLG. But there was a lot more going on, quite a bit of which, on the surface, doesn’t seem repeatable, that led to his outperforming his statcast numbers

I agreed with this at the time but re-stated my belief that “something is there.”

With a second season of statistics now available, how does Steven Kwan’s Hidden Power hold up to scrutiny?

Steven Kwan’s Hidden Power: 2023 Edition

First: know that Steven Kwan did not finish last in hard contact rate this season. He finished second to last. Follow the link if you do not already know the answer and do ponder at eleventh place.

I will now review the same numbers that I did last season.

Hold on. Nope, that’s a corgi and Steven Kwan smiling about ice cream. Or is that sherbet? Sorbet? Sure, Bert.

Anyway, here are his 2023 statistics compared to the League. SLG is Slugging Percentage, xSLG is Expected Slugging Percentage, and BABIP is Batting Average on Balls In Play.

Kwan, 2023 SLG: .370
MLB, 2023 SLG: .414

Kwan, 2023 xSLG: .358
MLB, 2023 xSLG: .413

Kwan, 2023 BABIP: .294
MLB, 2023 BABIP: .297

Kwan generated exactly 100 wRC+, a genuine league-average bat.

His BABIP regressed to about League Average rather than being .033 points above average as in 2022.

The difference between his xSLG and the League’s xSLG varied less. In 2023, it was .055 points while in 2022 it was .064.

Finally, Kwan lagged .044 points of SLG behind the league average in 2023 rather than slightly outpacing it as he did last year.

That the average between the difference in BABIP and the difference in xSLG happens to equal the difference in actual SLG is hilarious. There cannot possibly be anything there, but just in case, note that the equation “delta BABIP over delta xSLG equals delta SLG, baseball is solved dadgummit” was uttered in between giggles.

Coincidence acknowledged and discarded (?!), we can proceed.

Note that Kwan once again outperformed his xSLG. This is important, as an inability to do this at all might completely disprove the idea that sparked this line of thinking. If he did not, then it might be time to conclude that Steven Kwan is not able to outperform his xSLG due to something that advanced analytics misses. As the Statcast Glossary says, xSLG is useful because, “Expected Slugging Percentage is more indicative of a player’s skill than regular slugging percentage, as xSLG removes defense from the equation. Hitters, and likewise pitchers, are able to influence exit velocity and launch angle but have no control over what happens to a batted ball once it is put into play.”

Remember how I said I am making an argument at the edge of things? There is not anything in the above explanation that I do not agree with. Indeed, a batter who puts a ball into play cannot control what happens to that ball, but is it also true that the outcome of the batted-ball event cannot be influenced by coaching, individual hustle, or any other number of intangibles?

The conclusion I came to last time, again, was “maybe not!” and I extrapolated to suggest that perhaps Steven Kwan is just exceptional at rounding first base in such a way that allows him to turn a hit into a double that we would not otherwise expect just often enough that it shows up as the difference between his SLG and xSLG. The length of that sentence alone should indicate how likely it is to be plausible! But it was a place to start.

And yet, something that Steven Kwan continues to do resulted in a very slight difference between his SLG and xSLG in 2023. The difference is not as great — .012 in 2023 vs. .059 in 2022, but a difference remains. This is despite a slightly less-than-average BABIP and variations in the league-wide environment.

I also reviewed his “burst” last year. I’m going to use the same image again because you look at it

and then you understand “burst”. Let us look at Steven Kwan’s bursts, shall we?

Steven Kwan, 2022 Burst: 1.7, 6th among qualified
Steven Kwan, 2023 Burst: 0.9, 29th among qualified

Also of note is that Statcast indicates Kwan was slightly below average at reacting to batted balls and slightly above average at taking a route to intercept a batted ball. Last season, he came in at average for both. Exactly. I recall three seasons of defensive data being the rule of thumb for a suitable sample size, and it is interesting that for the second consecutive season, Kwan ranked as above average in this category.

It is worthwhile to repeat the math that SaladCzar offered last season. In 2023, Kwan had 236 total bases over 638 ABs for a .370 SLG. An xSLG of .358 would suggest his xTB would be ~228, or 8 fewer. Kwan had 43 2B + 3B this year. Based only on this season, we could not repeat their claim that “We can tell that the hidden power, should it exist, is not solely baserunning.” I still think that this is true, but it is interesting to see the numbers again.

Now, does Burst correlate to xSLG outperformance again? In 2023, The top 10 guys in Burst (Minimum PA: 10! I am not actually sure which subset of eligible players SaladCzar used last season!) outperformed their xSLG by an average of about .014 points, the top 25 by about .032 points, and the top 50 by about .026 points.

It appears that Burst may somewhat correlate to a player’s ability to outperform their expected slugging percentage for the second consecutive season. Neat!

We drew parallels between this stat and the “decision to round first” last time. While we’ve focused solely on Steven Kwan, we did mention Sandy Alomar, too. Coaching is critical in baserunning and so we will say “Sandy Alomar” again. How a player is specifically instructed to run the bases in different scenarios is largely invisible to me, but the difference we are seeing may also be impacted by in-the-moment coaching.

Truly, I do not know. I hope you found the process of checking these numbers once again to be as interesting as I did. I’m not sure if Kwan’s third season will provide more or less evidence in favor of an odd idea. Either way, it will be fun to find out!