clock menu more-arrow no yes

Filed under:

A Better Approach to Hard Hit Percentage, and Early Cardinal Results

New, 63 comments

Hard contact is nice, but it’s even better when it’s accompanied by fewer strikeouts, as the Cardinals have shown early this year

Milwaukee Brewers v St Louis Cardinals Photo by Jeff Curry/Getty Images

I’ve been trying to pay more attention to Statcast info this year, and use the info appropriately. It’s why I’m not necessarily going to flip out about someone’s expected wOBA after 12 games, but might emphasize max EV (exit velocity), which becomes meaningful quickly. One quandary came up recently regarding hard hit percentage. When writing about max EV during spring training, a reader contacted me to point out that while these metrics are interesting, they omit strikeouts. It’s a great point- barrel % and hard hit % are calculated by dividing by batted ball events. Lots of plate appearances end without batted balls. All things equal when comparing two players- same K%, same BB%- then you obviously want the hitter with the higher hard hit %. That scenario, however, doesn’t exist. Players have different K and BB rates. Pitcher List (an awesome site, by the way) went into great detail about all of this just last week.

There’s an easy solution offered in the Pitcher List article. Instead of calculating Hard Hit % as Hard Hits/Batted Ball Events (BBE), you can replace BBE with plate appearances. Then you have a more accurate picture of how frequently a hitter actually hits the ball hard. If you want a practical application of why this is important, let’s consider Nolan Arenado.

If we use the original Hard Hit % (HH%) formula, Arenado’s numbers have been:

2018: 40.2% (74th percentile in MLB)
2019: 37.6% (58th percentile)
2020: 33.7% (39th percentile)

It looks like he had a down year in 2020 and has been trending down for a few years. He actually did have a down year- I won’t sugarcoat it- but he also put the bat on the ball more frequently last year, not captured in HH%. His K% was the lowest of his career, and his contact% was the second highest of his career. Let’s reconfigure Arenado’s 2018-2020 using a true hard hit percentage (THH%- which I’m calling it; that’s not a formal name), divided by PA instead of BBE:

2018: 28.4% (83rd percentile in MLB)
2019: 28.5% (82nd percentile)
2020: 27.9% (76th percentile)

He slipped due to the shoulder injury in 2020, but not as much as HH% would have you believe. HH% had him in the lower half of the league. But by THH%, he was upper quartile. The decline that HH% saw from 2018 to 2019... actually wasn’t decline at all.

What can this tell us about the St. Louis Cardinals’ offensive performance thus far? I’ve collected every hitter in the league with over 100 PA in 2018 and 2019; over 50 PA in 2020; and 20 PA this year. Then I’ve calculated their THH%, and given them a percentile rank league-wide in both HH% and THH%.

Here’s how it looks for the Cardinals so far. You’ll see their HH% percentile, THH% percentile, and the difference between the two (THH-HH). Players with a large positive difference are underrated if you look strictly at HH%. Players with a large negative difference are overrated by HH%.

2021 Cardinals: Hard Hit % vs. True Hard Hit %

Name wRC+ Hard Hit Pctile THH Pctile THH - HH
Name wRC+ Hard Hit Pctile THH Pctile THH - HH
Tommy Edman 117 51.4% 80.3% 28.9%
Nolan Arenado 156 67.7% 78.4% 10.7%
Yadier Molina 159 59.4% 69.2% 9.9%
Paul Goldschmidt 75 96.9% 99.5% 2.6%
Dylan Carlson 138 25.0% 19.9% -5.1%
Paul DeJong 67 13.4% 7.0% -6.5%
Matt Carpenter 35 98.5% 83.3% -15.1%
Justin Williams 32 87.8% 68.4% -19.4%
Austin Dean 129 66.5% 43.9% -22.6%
Tyler O'Neill 27 90.0% 54.6% -35.4%

There’s Arenado again. His HH% has increased, but his THH% still does a better job of capturing his performance thus far. The only reason I left wRC+ in there is to show how much of an outlier Paul Goldschmidt’s performance has been. He’s consistently clobbering the ball- 48% of his plate appearances are resulting in a hard hit ball. That’s third best in the league through Wednesday. And yet, it’s only resulted in a 75 wRC+. Predictably, THH% thinks more highly of hitters who consistently put the bat on the ball (Tommy Edman, Yadier Molina) while thinking less of those with contact issues (Tyler O’Neill, Austin Dean). It’s not exactly a surprise, but still- it shows us that THH% is capturing a good part of what we hoped it would capture.

There’s an important element missing here, though. Dylan Carlson’s THH% is in the bottom quartile, as is Austin Dean’s, yet both have wRC+s well above average. Edman and Goldschmidt are upper quartile but are quite a bit beneath Carlson and Dean, particularly in Goldschmidt’s case. Arenado’s THH percentile is only slightly better than 2020, yet his wRC+ has more than doubled from 76 to 156.

Some of this can be explained by barrels per plate appearance (Brl/PA). It’s one thing to hit the ball hard (95+ mph exit velocity) frequently. It’s quite another to hit it with optimal EV and launch angle. Edman’s Brl/PA is only 9th percentile in the league. He’s hitting it hard plenty, but rarely optimally. Carlson is 66th percentile. He’s not hitting it hard that frequently, but he’s doing a decent job of maximizing his opportunities. Goldschmidt is still a mystery, with a solid 61st percentile. Maybe that explains some of why his wRC+ is lagging, but only a small amount. Long story short, if Goldschmidt keeps hitting the way he has thus far, he’s going to explode. If you take nothing else away from this article, make it that.

I have yet to mention Matt Carpenter because he’s the biggest mystery of all. His true hard hit percentile is 83rd- upper quartile. His barrels/PA percentile is 96th(!). Even if we want to revert to classic HH%, he’s 98th (!) percentile. Half of his plate appearances this season have ended either in a walk or a hard hit ball, which is 24th best in the league. My goodness, he’s doing all of the right things. But his wRC+ is 35, propped up by four walks, a bunt single, and one extra base hit. There’s probably a much larger article to be written here about Carpenter alone. My instinct is that, like Goldschmidt, he should explode soon if he can repeat this process. But I’m not convinced there isn’t something else going on that explains his underperformance.

Team-Wide Results

There’s one last thing I want to look at today. Specifically, how are the Cardinals doing on a team-wide basis? And how does that compare to past seasons? This data omits pitchers, as does everything I’ve written about so far:

Cardinal League-Wide Ranks, 2018-2021

Category 2018 2019 2020 2021
Category 2018 2019 2020 2021
THH% 25.44% 23.12% 22.43% 28.57%
THH Rk 8 26 28 6
Brl/PA% 5.06% 4.58% 4.48% 6.45%
Brl/PA Rk 6 13 24 8
THH+BB% 34.24% 32.62% 32.63% 38.67%
THH+BB Rk 10 22 22 4

I’d say I’m surprised by 2018, but don’t forget that’s the year Matt Carpenter went supernova and earned MVP consideration with tons of walks, hard hit balls, and optimal contact. More importantly, this year’s team has taken major strides all around, to the point that they’re top 10 so far in THH%, Brl/PA%, and THH+BB%. However, their wRC+ is a mediocre 99 and doesn’t reflect their gains.

Here’s why that’s significant. Here’s every team’s non-pitcher wRC+ from 2015 to 2019, plotted against their THH+BB%. Obviously, 2020 and 2021 are omitted because of sample size issues.

It comes out to an r-squared of .456. That’s a pretty good relationship- if you hit the ball hard a lot and walk a lot, or some combination of the two, you’re probably going to be very productive as a team. See that red dot on there? That’s the 2021 Cardinals through Wednesday. Small sample sizes, blah blah blah... but that’s a team doing a LOT of the right things at the plate, but without seeing the results you’d typically expect.

Just for fun, I threw BB%, THH%, and Brl/PA% into a regression blender to predict wRC+ for teams from 2015-2019. I get an r-squared of .50- not bad. Using the formula it spits out, you’d expect this Cardinal team to have a 117 wRC+. Compare that to the 99 that they have so far. There are many reasons this might be the case. April suppresses offense, there’s a new ball, and most of all, we’re talking about 12 games- ripe for this kind of odd variance.

Still, if they can keep hitting this way or even approximate it, there are a lot of runs coming.