Ladies and gentlemen, can I please have your attention. I’ve just been handed an urgent and horrifying news story. I need all of you to stop what you’re doing and listen.
Offense is down in Major League Baseball.
In 2014, teams averaged just 4.07 runs per game, a figure that has steadily decreased since 2006. One factor is improved pitching. In addition to throwing harder than ever before, pitchers have increased the number of strikeouts per nine innings while allowing fewer walks. Home runs have decreased by 23% over this period. Does this mean that power is down?
Figure 1: Changes in Runs, Home Runs, Strikeouts, and Walks between 2006 and 2014
Maybe hitters have just been unlucky and are hitting the ball hard in all the wrong places. Maybe they’ve been crushing the ball in Kauffman, and getting under it a bit too much in Coors.
To find out what luck’s got to do with it, let’s first look at Home Run Tracker from ESPN (courtesy of ESPN Sports & Information Group), which gives a measure of how fortunate a batter was to hit a HR by indicating how many parks it would have left. The “Parks” value takes the dimensions of each field, the temperature at the time it was hit, the wind conditions, and elevation all into account. This value can range from zero (it wouldn’t have been a HR in any stadium if it weren’t for the conditions) to 30 (it would have been a HR in any park). Maybe this will account for the decreasing HR totals that we’ve observed.
However, when we look at the data, we see the opposite is true. Over the past eight seasons, fewer HR would have left between 21 and 30 parks, while more would only have left between 1 and 10. So hitters have actually been slightly luckier in recent years!
Figure 2: Changes in Distribution for ESPN Home Run Tracker’s “Parks” Value (2007 to 2014)
Home Run Tracker also provides the true distance for nearly every HR hit. However, these data will suffer from some selection bias, as they’re only available for HRs, which, as we just saw, depend on several factors. In an attempt to overcome this, let’s look at the longest HRs hit since 2007. Figure 3 shows the average distance of the top 5% of HR by true distance for each season between 2007 and 2014. I’ve excluded inside the park HR and HR for which data were not available (such as those hit in the Tokyo Dome) from the base, as these vary in number from year to year. The slope of the trendline indicates the average distance of the top 5% of HR has decreased by nearly 0.7 ft each year.
Figure 3: 2014 Mean Distance of Top 5% of Longest Home Runs (2007 to 2014)
Another variable that’s available from Home Run Tracker is the speed of the ball off the bat (mph). This is likely a better measure of power as true distance is more affected by environmental factors such as wind, temperature, and altitude. Using the same methodology as above, let’s compare the top 5% of fastest HRs off the bat. Again, we see that velocity of the ball on HRs has decreased nearly 0.2 mph each year.
Figure 4: 2014 Mean Distance of Top 5% of Fastest Home Runs by speed off the bat (2007-2014)
However, home runs aren’t the best way to measure power. There’s just too much selection bias. Again, whether a ball leaves the yard depends on many factors, including Mike Trout.
A better metric to assess power is fly ball distance, including those that are caught, fall for hits, and clear the fence. For the rest of this post, I’ll refer to this type of contact as HR+FB. Baseball Heat Maps pulls these data from MLB GameDay, going back to 2007. You can read more about how this is done and some limitations of the data here. The data sets used for this post are the HR+FB distance leaderboard, which includes the 300 players with the most fly balls each season. These data are not without their own limitations, such as not having every HR+FB on record, and the distances for an out is recorded where the catch was made, rather than where the ball would have landed. However, a benefit is that infield fly balls are filtered out by using a minimum FB distance.
Over the period that these data have been collected, Jim Thome has the best season on record posting an average HR+FB distance of just over 323 ft. Michael Stanton clearly avoided the sophomore slump as his 2011 season ranks a close second at 322 ft. No, this isn’t the retired relief pitcher (though he did post a .735 OPS in 26 PA), Giancarlo’s name simply hasn’t been updated in the database. Either way, it’s not surprising that he’s up there as well.
Table 1: Top 10 Seasons for Average HR+FB distance (2007-2014)
Interestingly, though the data are available from 2007 to 2014, no season after 2011 is represented on the eight season leaderboard. Does this confirm a league wide decrease in power?
Before we measure the central tendency, we need to check if the distribution is normal. I’ve shown the distribution for the 2014 season below, using bins of 5 ft. However, the distribution is not normal for each season. This is likely due to the constraint of including only the top 300 hitters (by number of HR+FB),
Figure 5: 2014 Distribution of 2014 Average HR+FB Distance
I’ve chosen the median distance for players who qualified for the leaderboard as the measure of central tendency. Adding a linear trendline to the figure confirms that power has indeed decreased. The slope indicates a decrease of just over a foot each year. I also ran this after removing the median for 2009 and 2011, the two seasons that did not have a normal distribution (Shapiro-Wilk p-value < 0.01), which showed a slope of -0.929 (just under a foot per year).
Figure 6: Median HR+FB Distance for Top 300 Hitters by Number of HR+FB 2007 to 2014
So based on the evidence above, it appears that power in MLB has decreased! But this left me wondering, who hit the ball the furthest in the air relative to their peers each season since average HR+FB distance has been available?
To judge this, I created a metric that I’ll call Distance+ which compares the average distance for a single player to the weighted average distance of all players who appeared on the leaderboard for that year, the same principle as stats like OPS+. This is clearly less of a rigorous endeavor than the analysis above, but it’s still fun! Results are shown in Table 2. Rank corresponds to the overall ranking for all seasons, as shown in Table 1.
So how does sorting the eight season leaderboard by this variable change the results? For starters, on the strength of the rest of the league in 2010, Thome’s leading season drops to seventh on the Distance+ leaderboard and Stanton claims top spot. This leaderboard adds three new sluggers from 2011, 2013 and 2014, the three worst years on record for HR+FB distance years on record, while distance leaders not named Jim Thome from 2008-2010 fall out of the top 10. In these years of moderate power decreases, Paul Goldschmidt (2), Carlos Gonzalez, and Corey Hart retained their power better than their peers. Oh, and Jack Cust finds himself in the top 10. I guess that will happen when you hit 26 bombs while leading the league in walk (20.7%) and strikeout (32.3%) rate. That season began a three-year stretch of his role as three true outcomes archetype.
Table 2: Top 10 Seasons for Average HR+FB Distance+
|Name||Year||Hits||Distance||Leaderboard Avg||Distance +||Rank|
Finally, you can’t talk about power and ignore the players that displayed the least of it. So who had the worst power season according to Distance+? Jamey Carroll, of course! Minnesota Twins fans will likely remember his dreadful 2012 campaign. It would (rightfully) be his last as a regular. Fans of grit and hustle will recognize many of their heroes on this list as well.
Table 3: Worst 10 Seasons for Average HR+FB Distance+
|Name||Year||Hits||Distance||Lg Avg||Distance +|
You stay classy, baseball fans. I’m Matt Jackson?
HR data from ESPN Home Run Tracker.
HR+FB distances from Baseball Heat Maps.
All other data from Baseball Reference.
Next post: Eddy’s Baseball Autographs: Key to Success Is Finding the Right Card
Previous post: On Luis Valbuena, Whom You Shall Not Cast Aside So Easily