(*) RESEARCH: Home run streaks

Research

Patrick Davitt

Apr 24 2013 11:05pm

(*) RESEARCH: Home run streaks

Patrick Davitt

Apr 24 2013

The other day, on a telecast of a major-league baseball game, one of the commentators (an ex-player) noted, “We all know that those big home run hitters hit ’em in bunches.” He went on to say that the elite power hitters will often go two weeks without a tater before enjoying a homer spree.

At the risk of shooting fish in a barrel (ex-players being on the air chiefly to provide clichés and mostly uninteresting anecdotes, not insight), we thought these paired assertions deserved examination.

After all, if there’s something to this, we might be able to identify patterns that would help us stream batters, drop them, trade them (or for them) and generally improve our power counts in the fantasy games we play.

So we looked into it. And, sadly but unsurprisingly, the “homer bunch” theory holds about as much water as the “rising fastball” theory—which is to say, none at all.

Method

To test the hypothesis, we took every hitter who clouted 35 or more HR in any season 2010-2012. We identified 23:

Batter_Season       HR     Batter_Season       HR
==================  ==     ==================  ==
Bautista_2010       54     Hamilton_2012       43
Bautista_2011       43     Kemp_2011           39
Beltre_2012         36     Konerko_2010        39
Braun_2012          41     Pujols_2010         42
Cabrera_2010        38     Pujols_2011         37
Cabrera_2012        44     Reynolds_2011       37
Dunn_2010           38     Stanton_2012        37
Dunn_2012           41     Teixeira_2011       39
Encarnacion_2012    42     Uggla_2011          36
Fielder_2011        38     Votto_2010          37
Granderson_2011     41     Willingham_2012     35
Granderson_2012     43

Then we painstakingly logged every HR and all the batters’ games played, and parsed the resulting data to explore:

How often did these sluggers hit HRs in “bunches”?

How often did they experience long droughts?

How often were the droughts followed by bunches?

Were there any indicators to provide advance warning that either a drought or a bunch is coming?

One note about the results: Any time we describe a span of games between HRs, the count excludes the first game with a HR in the span. So two straight games with a HR have a gap of 1, not 0.

Note also that we defined a batter game as a game in which he actually played; if he sat out for any reason, that game is not counted in his streaks or droughts.

Results 1: Overall

The batters in the study averaged just over 4 games between HRs, in a range of averages from 3.5 to 5.0:

Batter_Season        G     Batter_Season        G
=================  ===     =================  ===
Stanton_2012       3.5     Willingham_2012    4.2
Bautista_2010      3.5     Votto_2010         4.3
Encarnacion_2012   3.6     Braun_2012         4.3
Hamilton_2012      3.7     Pujols_2010        4.4
Bautista_2011      3.8     Beltre_2012        4.4
Dunn_2012          3.9     Teixeira_2011      4.5
Granderson_2011    4.0     Cabrera_2010       4.5
Cabrera_2012       4.1     Fielder_2011       4.6
Konerko_2010       4.1     Reynolds_2011      4.8
Granderson_2012    4.1     Dunn_2010          5.0
Pujols_2011        4.1     Uggla_2011         5.0
Kemp_2011          4.2

Results 2: Bunches

The batters studied had 202 occurrences of hitting HRs in consecutive games. Two HR in two games seemed a little paltry to define a “bunch,” however, so we defined a “bunch” as hitting 5 or more HR in a 10-game span.

By this definition, the 23 batters had 22 “bunches” among them in their individual seasons, about 0.6% of all available 10-game spans. But only six of them had more than one bunch in a season:

Three bunches        One bunch           No bunches
=================    =================   =================
Bautista_2011        Beltre_2012         Braun_2012
Cabrera_2012         Encarnacion_2012    Cabrera_2010
                     Fielder_2011        Dunn_2010
Two bunches          Granderson_2011     Granderson_2012
=================    Hamilton_2012       Konerko_2010
Bautista_2010        Pujols_2010         Pujols_2011
Dunn_2012            Teixeira_2011       Reynolds_2011
Kemp_2011            Uggla_2011          Votto_2010
Stanton_2012                             Willingham_2012

José Bautista might be characterized as prone to homering in bunches, with five occurrences over his two 35+ HR seasons. But no other batter shows any evidence of unusual bunching.

If we reduce the requirements of a “bunch” to 3 or more HR in a six-game span, the number of bunches jumps to 155, but that is still less than 1% of the available three-game spans. Rare, in other words.

Results 3: Droughts

We faced a similar issue with defining a “drought.” Somewhat arbitrarily, but probably generously, this research defines a HR drought as a span of 10 games in which a particular player hits no home runs.

By this definition, these 23 batters endured 74 droughts in their high-HR seasons, barely 2% of all available 10-game spans.

On average, the studied batters had maximum gap-day occurrences around 16 days.

The largest gap between HRs was 28, by Albert Pujols in 2011 (Dan Uggla’s 24-game drought in 2011, and Giancarlo Stanton’s 20-gamer in 2012 were the only other occurrences of gaps greater than 20 games).

The smallest maximum HR gap was 11 days, by Curtis Granderson in 2012, with Ryan Braun and Josh Hamilton both having 12-game droughts in 2012.

In all, the players averaged about 3.2 droughts per year, ranging from five for Kemp_2010, Pujols_2010 and Teixeira_2011 to seven with two droughts apiece.

But this again suggests droughts are relatively rare. Even the high of five droughts in a season is well less than 1% of available 10-game spans, and the average of around three such droughts is likewise vanishingly rare.

Further, there is no evidence that once a hitter ends a drought, he begins a bunch. The average day-gap between the HR that ended a drought and the next HR was 3.7 games. Some players Braun, Encarnacion, Granderson and Stanton averaged 1.5 games or fewer, others were at 6.0 or higher. In 2010, Encarnacion had a 14-game drought, ended it with two swats, then had a second 11-game drought.

In all these seasons and spans, only one hitter followed a drought with a bunch. Giancarlo Stanton went through a 20-game drought near the start of 2012, then hit five HR in his next eight games. Also in 2012, Hamilton had the reverse happen: He had eight HR in five games (including a four-tater eruption in one game) , then had a 12-game drought.

The idea that these elite HR achievers follow any kind of drought-bunch pattern is simply not borne out by the evidence.

Results 4: Is There Any Pattern?

No.

Conclusion

Notwithstanding the expert analysis of the former players who grace the sport’s broadcast booths, there is no evidence whatever that HR hitters operate in bust-and-boom cycles.

For that reason, it is a mug’s game to try to predict either the beginning or the end of either a drought or a bunch, or to assume the end of one presages the beginning of the other.

Simply put, HR hitters hit HRs in what is a random walk, with day-gaps between HR that correspond roughly to their average days per HR. A 27-HR guy will hit one HR about every six games, with random variation to either side of that mean. If he hasn’t hit one for eight games, he is still not any more (or less) likely to hit one in the ninth, or the tenth. He will have a few streaks where he goes without longer than expected, and a few streaks where he hits more frequently than expected.

Now, what’s all this about rising fastballs and balls that “pick up speed” after they’re hit?