Getting to Know Barrels Per True Fly Ball Rate (Brls/True FB)
Since it was first introduced about a year and a half ago and published on the Statcast Leaderboard, Barrels quickly became one of my favorite metrics. Combining exit velocity with launch angle made it the perfect statistic to reference when investigating a hitter’s power potential. Since we like ratios better than counting stats for projection purposes, Barrels per Batted Ball Event (Brls/BBE) was my metric of choice. But as informative as it remains, I discovered that it actually wasn’t the best ratio when it came to forecasting HR/FB rates.
Yesterday, I unveiled the newest version of my batter xHR/FB rate, which looks a bit different than my original Statcast-charged version. One of the changes made to the equation was a switch from using Brls/BBE to Brls/True FB. That is, instead of using all batted balls as the denominator, I realized that we really only care about fly balls, since that’s the denominator of the equation we’re trying to update. Duh. But rather than use all fly balls, I removed pop-ups, to get a truer sense of a batter’s ability to hit his fly balls over the fence. I dubbed this statistic True FB.
Since Brls/True FB (Brls/TFB from now on) is a new metric I just introduced yesterday, let’s get to know it, shall we? To put it into context, I will do a lot of comparing to Brls/BBE. Note that all statistics are solely from the population I used to create my equation, which are batters with at least 60 BBE from 2015 to 2017 (1,347 total over three seasons).
| Season | Population Avg Brls/TFB | Population Avg Brls/BBE |
|---|---|---|
| 2015 | 17.99% | 5.48% |
| 2016 | 20.45% | 6.37% |
| 2017 | 19.94% | 6.41% |
While my population’s average Brls/BBE spiked in 2016 and rose again ever so slightly in 2017, Brls/TFB actually regressed marginally in 2017 after surging in 2016. That’s probably a direct result of more fly balls hit in 2017, providing more chances for batted balls to be classified as a barrel. But when we’re isolating fly balls, we find that hitters aren’t barreling them up more often, which might come as a surprise given the league HR/FB rate trend.
| 2015-2017 | Brls/TFB | Brls/BBE |
|---|---|---|
| Max | 63.5% | 25.7% |
| Average | 19.5% | 6.1% |
| Median | 18.2% | 5.9% |
Since TFB is a much smaller subset of BBE, Brls/TFB is of course going to be a larger rate than Brls/BBE. Because of the bigger maximum, the range is a lot wider, increasing the skill differences between players. Can you guess who easily led baseball over the last three seasons with the max in both metrics? Aaron Judge, obviously.
This table helps us determine what’s a strong, average, and weak Brls/TFB mark, something that’s important to remember when learning a new metric. My population average and median Brls/TFB rates are about triple the marks of Brls/BBE. Though, the max Brls/TFB isn’t quite triple that of Brls/BBE (c’mon Judge, get on that!).
Now let’s compare where a hitter ranked in 2017 in Brls/BBE to where he ranks in the new Brls/TFB. I subtracted the Brls/BBE rank from the Brls/TFB rank, so a positive number mean a worse Brls/TFB ranking, while a negative number means better. Let’s start with those who look less powerful using Brls/TFB.
| Player | Brls/True FB | Brls/True FB Rk | Brls/BBE | Brls/BBE Rk | Rk Diff |
|---|---|---|---|---|---|
| Kyle Seager | 17.8% | 219 | 8.6% | 117 | 102 |
| Matt Carpenter | 17.2% | 229 | 8.2% | 131 | 98 |
| Tyler Moore | 17.7% | 222 | 8.1% | 133 | 89 |
| Justin Turner | 18.5% | 203 | 8.6% | 116 | 87 |
| Gregory Bird | 18.4% | 206 | 8.4% | 127 | 79 |
| Pat Valaika | 15.3% | 255 | 6.7% | 176 | 79 |
| Lucas Duda | 26.3% | 109 | 12.1% | 33 | 76 |
| Anthony Rendon | 15.0% | 262 | 6.5% | 186 | 76 |
| John Jaso | 18.1% | 211 | 7.9% | 138 | 73 |
| Josh Phegley | 12.7% | 293 | 5.6% | 222 | 71 |
| Jay Bruce | 23.8% | 132 | 10.5% | 62 | 70 |
| Derek Norris | 12.3% | 306 | 5.2% | 236 | 70 |
These hitters were really crushed with the switch to Brls/TFB. Why? Because as a group, they averaged a whopping 45.8% True FB% versus just a 31.6% population average mark from 2015-2017! Their fly ball happiness boosted their Brls/BBE marks thanks to all the extra opportunities they had to hit a barrel.
Now let’s check on the winners of this metric. You could probably guess that it’s the low True FB% crew.
| Player | Brls/True FB | Brls/True FB Rk | Brls/BBE | Brls/BBE Rk | Rk Diff |
|---|---|---|---|---|---|
| Derek Fisher | 31.3% | 62 | 5.4% | 228 | -166 |
| Jonathan Villar | 30.0% | 72 | 5.5% | 223 | -151 |
| Yandy Diaz | 19.0% | 195 | 3.3% | 321 | -126 |
| Howie Kendrick | 40.4% | 12 | 7.9% | 135 | -123 |
| Eric Hosmer | 34.0% | 41 | 7.0% | 162 | -121 |
| David Freese | 29.4% | 77 | 6.4% | 187 | -110 |
| Christian Arroyo | 16.7% | 236 | 3.2% | 326 | -90 |
| Erik Gonzalez | 46.7% | 5 | 9.3% | 90 | -85 |
| Ian Desmond | 14.6% | 267 | 2.7% | 352 | -85 |
| Craig Gentry | 21.1% | 167 | 4.9% | 247 | -80 |
| Ryan Rua | 36.8% | 24 | 9.0% | 103 | -79 |
| Russell Martin | 29.1% | 80 | 7.1% | 157 | -77 |
| Christian Yelich | 28.7% | 86 | 7.0% | 163 | -77 |
| Yunel Escobar | 21.4% | 164 | 5.0% | 241 | -77 |
| Joe Mauer | 19.8% | 188 | 4.5% | 263 | -75 |
| Leury Garcia | 18.9% | 200 | 4.2% | 275 | -75 |
| Tommy Pham | 38.8% | 19 | 9.3% | 91 | -72 |
We were right! This group’s True FB% was a measly 22%, which gave them scarce opportunities to hit a barrel. So all this time, the Brls/BBE underrated their home run power potential. This is mostly a list of scrubs, certainly a less impressive group of hitters, but there are a couple of intriguing names.
Derek Fisher is a great name to tuck away if he ever managed to fall into playing time on that ridiculously good Astros team. Or hey, perhaps they trade him. Check out Howie Kendrick! From a non-descript single digit Brls/BBE to a monstrous 12th ranked Brls/TFB! Eric Hosmer’s name is no surprise to find, as he is the king of hitting too many grounders for a guy we know has good power. Ummm, Erik Gonzalez fifth? WHO?! Pretty small sample size, but the power breakout first began at Triple-A, suggesting that this was at least somewhat for real.
I purposely lengthened the list so I could grab Tommy Pham at the end of it. It’s not often that a 29-year-old career minor leaguer records his first season with more than 200 MLB plate appearances and posts a .398 wOBA, .214 ISO, and 26.7% HR/FB rate. With just a 9.3% Brls/BBE mark that ranked 91st, it was impossible to believe that mid-20% HR/FB rate was anything close to sustainable. But a 38.8% Brls/TFB rate that ranked 19th? Yeah, that offers a heck of a lot more reason to be optimistic that this wasn’t just a dream fluke for Pham. And since I know you’ll ask, his updated xHR/FB rate was 22.5%, significantly higher than the 15.2% xHR/FB rate calculated from the previous equation.
More Brls/TFB rate talk to come!
Mike Podhorzer is the founder of ProjectingX IQ, an advanced fantasy baseball analytics platform that transforms projection data and in-season performance signals into actionable intelligence. He is the 2015 Fantasy Sports Writers Association Baseball Writer of the Year and three-time Tout Wars champion. He is the author of the eBook Projecting X 2.0: How to Forecast Baseball Player Performance, which teaches you how to project players yourself. Follow Mike on X@MikePodhorzer and contact him via email.
Okay, interesting. So would you say both of these statements are equally accurate/fantasy-relevant?
1) “The bottom list is populated by guys who have shown the potential for a power breakout if they start lifting the ball more.”
2) “The bottom list is populated by guys who are already turning their relatively few fly balls into barrels at an unusually high rate – if they don’t/can’t adjust to hit more fly balls, they have likely maxed out their possible power production.”
Thanks!
Kind of to number 1. The bottom list is a group of guys my original equation underrated for HR/FB forecasts. That doesn’t mean they have greater breakout potential. However, because their FB% marks are low, it seemingly gives them more room for growth, which would indeed lead to a breakout. That said, any hitter who increases their FB%, with all else remaining equal, will enjoy a homer spike.
If you phrase it in such a way to state that Giancarlo Stanton and Aaron Judge have maxed out their possible power production, then 2 is accurate. Obviously a high HR/FB rate is harder to maintain and the powers of regression and effects of aging are going to weight on that mark. But again, it’s no different than any group of hitters with a high HR/FB rate, remaining in the top quadrant in a skill is hard!
signs of hope for ian desmond? the GB rate went way up last year
fascinating
Hmm between this piece and the one on the front page about Moran, makes me wish the Pirates had gotten Fisher and Moran for Cole — a not astronomical (it’s not Tucker or Whitley) but exciting package.
This is excellent work Mr. Podhorzer. Do you know how correlated Brls/TFB is year to year? I would expect it to be similar but slightly less than Brls/BBE.
I meant to calculate some correlations and compare to Brls/BBE. Too much to write about, wasn’t sure what to include in my first article about the metric! Will look into doing it in tomorrow’s post.
Do any of these metrics correlate year-over-year? Do they stabilize after a certain number of BBE or “True FB”? Do they correlate with future HR per PA? (I say per PA because HR per “True FB” doesn’t matter as much if the player doesn’t hit a lot of “True FB”.)
Correlations will likely come tomorrow. I don’t know how to calculate the stabilization point. Help?
HR/True FB matters just as much as HR/FB for ground ball guys. Sure, a 20% jump in either metric will add fewer homers than a fly ball guy, but it still has the same impact. But maybe that’s what you meant. Going from 10 to 11 homers isn’t as valuable as going from 30 to 33, even though both are 10% increases.
https://www.fangraphs.com/library/principles/sample-size/
>But when we’re isolating fly balls, we find that hitters aren’t barreling them up more often, which might come as a surprise given the league HR/FB rate trend.
The way I read this is, you have your Yonder Alonso, and you will also have Trevor Story. Previously we look at Brl/BBE because Brl is what we drew as metrics to producing HRs. The new equation didn’t change Brl, just highlights the guys who are hitting ball hard but not hit _enought_ flyball.
In my mind, Group 1 is already there. Maybe they don’t _always_ square it up when it comes to flyball, but their Brl show you they belong.(at least I care about top 150 in raw Brl). Group 2 is the audience you can preach lifting the ball. Hopefully if they do, you can see them hit more HRs.
I am intrigued in what Daniel Murphy will tell his teammate Howie Kendrick if he read this article.
Yes, you’re correct. But as I commented above, any hitter, no matter their current FB%, is going to hit more homers if they increase their FB%, all else being equal. But of course the low FB% has seemingly more room for growth.
Interesting article. I have some concerns about the validity of removing pop-ups from TFBs. I absolutely understand removing line drives and ground balls from BBE to calculate TFBs. However, when you look at the two groups above the over vs under performers in the two charts above I would be willing to hazard a guess that the spread between the two groups is somewhat minimized. With group that hits a higher rate of fly balls likely being more pop-up prone than the group that has the lower fly ball rate having a lower instance of pop-ups. In the end the effectiveness of stat boils down to exactly what you’re trying to calculate. If you’re trying to get a good sense of how many home runs an individual player may hit in a year, I would say that the new metric isn’t the most effective but, if you are trying to get a sense of a players home run potential then the new metric is probably fairly accurate, with the caveat that players in the second chart likely won’t reach their theoretical home run potential as any significant increase in the rate if TFBs will likely come with a corresponding increase in pop-up rate. However, if pop-up to TFB rates are relatively predictable it should really only be a matter of solving a relatively simple optimization problem to determine any given players “ideal” TFB rate and “peak xHR/FB rate.”
First off, pop-ups are a separate batted ball type, and should never have been lumped in with fly balls to begin with. They have different BABIP marks and expected results, so they should have always been a separate batted ball type like a line drive.
Next, it doesn’t actually make a difference for xHR/FB rate, because I’m ultimately adding the pop-ups back in to arrive at the expected mark, which is directly comparable to the hitter’s actual HR/FB rate. Removing the pop-ups allows us to determine whether a hitter’s HR/FB rate dipped because of a higher pop-up rate, or because they were actually hitting fewer of their fly balls over the fence. I want to separate the two, because to me, they are two different skills. It means more work for me in the future though, because now I have to project pop-up rate! But if it means slightly more accurate projections, so be it.
Thanks for the response. Your new need to project pop-up rate is basically what I was attempting to get at with part of my comment. The other was more wondering exactly what you saw as the use case for the statistic, which I believe you also addressed.
Wonderful. Just top notch stuff Pod. Gonna have to nommo this series for fantasy writer awards when the time comes. Please post the full list of xHR/FB? Please (scratching forearm furiously).
Thanks! Not sure I’ll post the entire list, but I’ll have my usual series of upside/downside guys, surprises at the top/bottom, validations (like Tommy Pham), etc.
Mike good data here. Curious where Joey Gallo fell in this ranking. Thx
Gallo ranked 2nd in Brls/BBE, tied for 5th in Brls/True FB. He ranked 4th on my file in True FB%, so that’s why he ranked slightly worse in Brls/True FB. So many opps for a barrel!
Love it, thanks Mike. This is why I subscribe!
In the first table, the yearly averages for Brls/TFB and Brls/BBE are all greater than the combined averages in the table below it. What am I missing? Maybe one table uses a weighted average and the other doesn’t?
Because I’m an idiot 🙂 Instead of summing up the raw numbers and then calculating the average from the totals, I stupidly just went and used the =AVERAGE formula on Excel on the percentages themselves…which, of course, made it an unweighted average, AKA, not actually an average at all. I did it because I was in that mode, having used the =MAX and =MEDIAN formula for that table too. I fixed it, thanks for the heads up!
I created a similar stat a couple months ago, called it effective off the ground percentage (100-GB%-PU%)
https://www.fangraphs.com/community/better-stats-for-finding-the-next-rhys-hoskins/
It includes line drives though. Does your stat also include line drives or just fly balls above 25 or so?