/cdn.vox-cdn.com/uploads/chorus_image/image/4132999/20120421_jla_aw4_817.0.jpg)
When a player has a breakout home run season at the plate, everyone is interested in analyzing whether the new found power can be sustained into the following season. Likewise, as a player ages, if home run totals drop for a season we sometimes wonder whether his power is truly diminishing or whether to expect a one year valley.
Some metrics that are available that I have both seen used by others and used myself in an effort to form a logical analysis of the potential power of a player include average home run distance as well as average home run plus fly ball distance. It would make sense to me that a player that hits the ball farther, on average, would possess more power and hence could be expected to produce more offensive "power" in the near future. I have never actually seen a study that has tried to investigate just how useful these logical indicators are in actually describing or predicting power.
This study will look at three potential predictors for two outcomes that describe power: HR and ISO. The three predictors are the two mentioned above as well as maximum home run distance. Given that the data involved for this study had to be merged from three sources, I have only examined 2011 and 2012 in this study.
Descriptive Capability
First, I ran linear regressions for all six combinations where all data was from the same season. This will show how well (or not so well) these potential indicators of power describe players' HR totals and Isolated Power in a given season.
I used all players that hit at least one home run in a given season for the sample set. The R-squared values were as follows:
2011 | HR | ISO |
---|---|---|
HR Avg Distance | 0.05 | 0.06 |
HR Max Distance | 0.39 | 0.20 |
HR+FB Avg Distance | 0.09 | 0.20 |
2012 | HR | ISO |
---|---|---|
HR Avg Distance | 0.05 | 0.06 |
HR Max Distance | 0.43 | 0.20 |
HR+FB Avg Distance | 0.09 | 0.32 |
The results are fairly consistent between the two seasons.
Average home run distance appears to be a poor indicator for actual home run totals as well as Isolated Power.
Average home run plus fly ball distance is also poor in describing variance in HR totals, but does appear to account for more of the variance in ISO. Logically I can imagine that on average longer fly balls that stay in the park are going to lead to more extra base hits than shorter fly balls.
Interestingly, the maximum home run distance achieved by players in a given season describes by far the most of the variance in the home run totals players amass in that same season. This is really a "raw power" metric, showing how far a batter can hit the ball when everything goes right.
Predictive Capability
With the data for two seasons in hand, I also tried running linear regressions for all six combinations but this time with the three predictor data sets from 2011 and the HR and ISO values from 2012. This is likely what we are most interested in knowing in attempting to use any of these metrics to determine what power levels we may be able to expect in the coming season.
I used all players that hit at least one home run in a given season for the sample set. The R-squared values were as follows:
2011 --> 2012 | HR | ISO |
---|---|---|
HR Avg Distance | 0.09 | 0.04 |
HR Max Distance | 0.27 | 0.13 |
HR+FB Avg Distance | 0.11 | 0.07 |
As predictors, none of these are particularly good. As points of reference, running a linear regression of 2011 HR totals and 2012 HR totals, the R-squared value was 0.35. From other research by Bill Petti we know in general ISO is very repeatable, with a year-over-year R-squared value of 0.72 between 2002-2009.
I tried a multiple regression of all three predictors to see if that improved the correlation with home runs. It was only slightly better than using just maximum home run distance, with an R-squared of 0.28. Both average and max home run distances were statistically significant at the 95% confidence interval in this experiment.
A multiple regression with ISO gave an R-squared of 0.16, with all three factors statistically significant at the 95% confidence interval. That said, the combination still accounts for a relatively small amount of the variance in ISO.
So it seems the single metric out of the three studied that appears to be somewhat worthwhile is maximum home run distance as a predictive factor for home runs. This could perhaps be looked at in conjunction with prior year HR totals in trying to determine expectations for HR totals in the following season.
To be honest this finding is surprising to me, in that average distances do not appear to be very useful gauges of power.
Interesting Cases from 2012 Heading into 2013
Here is a list of players from 2012 who had the highest maximum home run distances of all hitters that hit less than 25 home runs:
Name | HR | ISO | HR Max Distance |
---|---|---|---|
Cameron Maybin | 8 | .107 | 485 |
Nelson Cruz | 24 | .200 | 484 |
Travis Hafner | 12 | .210 | 481 |
Justin Maxwell | 18 | .232 | 471 |
Michael Morse | 18 | .180 | 465 |
David Freese | 20 | .174 | 464 |
Mike Moustakas | 20 | .171 | 464 |
Yoenis Cespedes | 23 | .214 | 462 |
Tyler Moore | 10 | .250 | 462 |
Rickie Weeks | 21 | .170 | 462 |
Andrew Brown | 5 | .196 | 460 |
Brett Lawrie | 11 | .132 | 459 |
It is not surprising to see guys that we think of as having significant power but missed part of the season due to injury make this list. Hafner, Morse and Cespedes are among those with this profile. Maybin's home park is deadly for home runs, although more so for left-handed hitters. Moustakas and Lawrie are young AL third basemen with high expectations from their fan bases. It will be interesting to watch how the Nationals handle Moore after the recent acquisition of Denard Span gives then another outfielder.
On the other end of the spectrum, here are players from 2012 who had the lowest maximum home run distances of all hitters that hit at least 10 home runs:
Name | HR | ISO | HR Max Distance |
---|---|---|---|
Norichika Aoki | 10 | .144 | 374 |
Coco Crisp | 11 | .158 | 402 |
John Mayberry Jr. | 14 | .150 | 405 |
Zack Cozart | 15 | .153 | 409 |
Alberto Callaspo | 10 | .109 | 411 |
Nick Markakis | 13 | .174 | 412 |
Shelley Duncan | 11 | .185 | 413 |
Omar Infante | 12 | .144 | 413 |
Dexter Fowler | 13 | .174 | 416 |
Desmond Jennings | 13 | .143 | 417 |
Will Middlebrooks | 14 | .221 | 417 |
Martin Prado | 10 | .136 | 417 |
Aoki barely made double-digits in the home run column in 2012, helped by one inside-the-park job. Markakis is an interesting name on the list, although he did miss time and likely was playing with a sore hand for another stretch of the season. Red Sox Nation has high hopes for their young third baseman as well, so for their sake hopefully an appearance on this list does not indicate that a power regression is in store from Middlebrooks.
What do you think? Do these results surprise you in any way? Would you still look to make use of any of these three metrics in analysis looking at a player's potential power production going forward?
You can follow me on Twitter at @MLBPlayerAnalys. <a href="https://twitter.com/MLBPlayerAnalys" class="twitter-follow-button" data-show-count="false">Follow @MLBPlayerAnalys</a>
<script>!function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0];if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src="//platform.twitter.com/widgets.js";fjs.parentNode.insertBefore(js,fjs);}}(document,"script","twitter-wjs");</script>
Credit and thanks to Jeff Zimmerman at Baseball Heat Maps, ESPN Stats & Information Group and Fangraphs for data upon which this analysis was based.