When looking at a baseball players stats, you will see a home runs column. Be that for batters or pitcher. When you see those columns, those homers are treated as equals. Even for the more advance stats like FIP, wOBA and my RFIP, from my previous piece, they all treat home runs as equals. With these advanced stats, all home runs are given a weighting and included into the metric. But are all home runs really equal?
Watch the following 4 home runs do you think these are equals?
Victor Martinez (2016-08-14)
Jedd Gyoko (2017-08-17)
Anthony Rizzo (2017-07-15)
C.J. Cron (2017-09-20)
I am hoping your answer is no, as that is my answer. To me the first two are quite lucky, dependent on the ballpark and proximity to the foul line. Where as the second two are ‘no doubters’ they would be homers in every stadium and in any fair direction. What if I told you that the hit probability of the first two were 0.4% and 0.5% but they second two were both 100%, does that sound reasonable to you?
Well according to MLB Statcast they were. They generate this value using a smoothing model built around launch angle and launch speed of batted balls, using these two factors they calculate a hit probability based on historic hits similar to that. MLB has captured this data the last four seasons since the start of 2015 and in that time period there have been over 500k batted balls. That data has given the model shown below.
This model then allows us to look at all the home runs that have been hit in the last 4 seasons and say what is the likelihood that they would be a hit. The histogram below shows the distribution of the hit probability of all 22,199 regular season home runs in the last 4 seasons.
As you can see the main volume is to the higher probability end but there are 4,405 home runs hit where the probability of a hit was less than 50%. That accounts for 19.8% of all home runs. When I selected the 4 home runs to show you, I selected the two with the lowest hit probability and two with the highest. Using the data below for those home runs you can see where they are in the model.
Victor Martinez (2016-08-14) – Launch Angle = 43.6°, Launch Speed = 93.1 mph, Distance = 325 ft
Jedd Gyoko (2017-08-17) – Launch Angle = 45.3°, Launch Speed = 95.0 mph, Distance = 346 ft
Anthony Rizzo (2017-07-15) – Launch Angle = 25.6°, Launch Speed = 109.5 mph, Distance = 445 ft
C.J. Cron (2017-09-20) – Launch Angle = 26.2°, Launch Speed = 109.9 mph, Distance = 437 ft
Given that we can quantify how likely a home run was, should we give the batter full credit for hitting home run when it had a low hit probability? Should we punish a pitcher the same for allowing a low hit percentage homer compared to a high percentage one? Personally I don’t think so and therefore the next step is determining which ones do we credit/punish.
Thankfully MLB Statcast may be here to help once again with a metric called Barrels. The Barrel classification was designed for batted-ball events whose comparable hit types (in terms of exit velocity and launch angle) have led to a minimum .500 batting average and 1.500 slugging percentage since Statcast was implemented Major League wide in 2015. Using that they came up with the following definition.
To be Barreled, a batted ball requires an exit velocity of at least 98 mph. At that speed, balls struck with a launch angle between 26-30 degrees always garner Barreled classification. For every mph over 98, the range of launch angles expands another two to three degrees until an exit velocity of 116 mph is reached. At that threshold, the Barreled designation is assigned to any ball with a launch angle between eight and 50 degrees.
So how does a Barreled hit compared to a non Barreled hit?
Barrels accounted for 5.4% of all hits across the four seasons with a batting average of .824 and home run rate of 60.4%. Compare that to a non Barrel and you see a batting average improvement of 260% and home run rate improvement of 5300%, yes that is correct a Barrel is 54 times likely to be a home run than a non Barrel. So let’s go back to our histogram for home runs but split out for Barrels and Non Barrels.
As you can see the low probability home runs were all non Barrels, armed with this I believe it would be good to credit hitters who can hit them and punish pitchers who give them up. The next question is how reliable are pitcher and hitter Barrel rates and how does that compare to home runs?
The table above shows the season to season reliability, using Pearson correlation, of Barrels and home runs by batted ball event (data for 2015-2018). Barrels has a greater reliability for hitters but has a very similar correlation for pitchers. To be honest I was hoping for an increased reliability for the pitching side as I was looking to use barrel rate instead of home run rate in my RFIP metric. I will still continue to investigate this but to work out the average run value of a barrel but I suspect that it won’t improve my system and if it does it would be minimal.
So that is all doesn’t end on a disappointing development let me tell you about the Baseball Savant Statcast Search, this is a great service which you can search with ease for specific pitches/outcomes and for most of them get a video of it. If there is an at-bat or pitch that you want to find the video for this is a great tool for you to use.
And now you can find with ease the most surreal moment of the 2018 season, a pitcher hitting a home run of a batter. Enjoy.
All videos and photos are copyright of MLB and were found using the Baseball Savant Statcast Search.