It is likely that you have heard about the ongoing strikeout scourge that is ravaging offense in Major League ballparks. League strikeout rate (K%) has been steadily rising over the last decade, and is up almost three and a half percentage points from the rate seen in 2006. The 20.4 percent rate posted in 2014 is actually the highest league strikeout rate ever; well, since 1916, the first year for which it is available on FanGraphs. This season, each team is roughly 90 games into their campaign, and as a group they are striking out at a rate only a small fraction lower than that of last season. The current 20.1 percent rate is the second-highest rate seen since 1916. That is a lot of batters striking out, perhaps too many.
To go along with the striking out, batters are also not walking as often as they have in previous seasons. Walk rate (BB%) has declined almost one and a half percent from the early part of the last decade to the current low mark of 7.5 percent. That rate is actually tied for the fourth-lowest rate since 1916, mixed in with seasons from the Dead Ball Era (e.g., the 1917 - 1922 seasons). It is slightly lower than last season's rate of 7.6 percent, and prior to that we have not seen a rate that low since 1968 (also 7.6 percent). Taking the strikeout and walk rates together, we are in the midst of a historically interesting season.
The reasons underlying the change in strikeout and walk rates are well documented. Jon Roegele has done incredible work tracking the ever expanding size of the strike zone, and Streve Treder has an excellent article in the Hardball Times Annual (and online) detailing the history of the size of the strike zone. So team's should be aware of what is going on, and can perhaps plan their roster's accordingly. After the Red Sox signed (or traded for) a suite of ground ball pitchers who, at the time, consistently threw pitches in the bottom part of the zone - the area that has seen the largest expansion - there was suggestion that the front office had made the moves with the intention of maximizing any advantage that could be extracted from the new larger strike zone. The decisions to acquire free swingers Pablo Sandoval, Hanley Ramirez and Rusney Castillo can also be seen as a response to the way the strike zone is being called into today's game.
In any case, we have an interesting situation. Players are striking out more often than has ever been seen, and walking at a historically low rate. The game done changed. Teams that can take advantage of the strikeout on defense while also avoiding the strikeout on offense should be more successful. This is something that I looked into in May of last year, and wanted to follow up on. Specifically, which teams have been the best at being on the positive side of the strikeout scourge? And, second, to what extent does it relate to winning?
I collected team strikeout and walk data for the 2004 to 2015 seasons (through Tuesday's games). There was no scientific reason for selecting these seasons, other than ensuring many years of data and including years with a range of strikeout and walk rates. For each season, counts of strikeouts and walks (unintentional and intentional) were used to calculate a plus-minus value similar to that which I used in my previous article on this topic. The value is the net number of strikeouts and walks recorded on defense added to the net number of strikeouts and walks recorded on offense as follows:
PlusMinus = (Pitching K - Pitching BB) + (Batting BB - Batting K)
From a defensive perspective strikeouts are good and walks are bad, while the opposite is true from the offensive side of things. Thus, a positive PlusMinus indicates that overall the team has been on the right side of all these strikeouts.
To begin, here are the results for the 2015 season (through July 21st):
|Team||G ||Win%||pK||pBB||bBB||bK||Def||Off||+ / -
Interestingly enough, this simple calculation does a decent job of separating the teams that have fared well this season from those that have not. There are currently 15 teams that have a winning percentage of .500 or better, and of those nine teams have a positive PlusMinus score. Still, three of the top five (Indians, Athletics, and Red Sox) have losing records. Maybe this presents evidence of the Red Sox 'plan' working, though unfortunately it has not translated into wins in what has become a really ugly season. Seven of the ten current playoff teams are on the positive side of things, with the Astros, Twins, and Pirates the holdouts on the negative side. Even though it does a fair job of categorizing teams, the correlation between this PlusMinus number and winning percentage is only r = .351, meaning only 12% of the variance in winning can be accounted for by this simple measure. Exchanging actual winning percentage for Pythagorean winning percentage (as given on Baseball-Reference) does not yield a difference in results (r = .328). Although this makes some sense, as there are teams significantly underperforming their run differential (hello Athletics!), and there are teams overperforming to similar extents. Also note that these correlations are only based on a little over half a season (~90 games per team).
Looking at complete season data reveals that this simple measure holds reasonable explanatory power, with interesting season-to-season variation. Below is a table with the correlation coefficient for the correlations between PlusMinus and winning percentage, and PlusMinus and Pythagorean winning percentage across the last 11 seasons:
|Season||r with W%||r with pW%|
You can see that there is little difference in the correlations between winning percentage and pythagorean winning percentage. However, you can see that last season, when strikeout rate was the highest since 1916 and walk rate was bottom ten since 1916, had the lowest correlation. Perhaps the rates have gotten too high (and low) to be able to extract as much from team-to-team variation. Yet, in 2013, when K% (19.9) and BB% (7.9%) were not all that different from 2014, we see the second strongest relationship with winning percentage. So it is difficult to know what to take from this other than that there is a decent relationship here.
It seems unlikely that the trend in strikeouts will increase much from the present rate. The Commissioner is interested in attracting fans to the game and will certainly make efforts to reverse the trend. For this reason it is probably not a good idea to focus too much long-term team building initiative around exploiting the expansive strike zone issue. With that said, the analysis here shows that there is value in being on the right side of the strikeout scourge. Given the current state of the game, I thought that would be true now more than ever, but the data do not entirely support that contention.
. . .