June 23, 2010
Prospectus Hit and Run
Consider The K
Last week, I explored the frequently-voiced claim that we're experiencing another Year of the Pitcher. Digging into the numbers, I noted that while scoring was down about three percent from 2009 and had dipped to its lowest rate since 1993, a more intriguing facet of the current crop o' stats is that strikeout rates are at an all-time high. This year, batters are whiffing in 18.1 percent of all plate appearances, up one tenth of a percent from last year, over one percent higher than 2007, and over two percent higher than 1994. Where are the strikeouts coming from?
While scoring has been all over the map over the past two decades, strikeout rates have crept upwards at an increasingly accelerated pace. Examining the annual pitch data at Baseball-Reference.com, which goes back to 1988, yields some subtle insights. The full data culled from B-R is here, but allow me to summarize:
I've isolated two time spans here. The first one runs from 1990 through Sunday, a range chosen to provide a sample which includes pre-expansion, pre-strike data. The other runs from 2000 through 2009, to sidestep the 1994-1995 strike and the currently incomplete campaign in favor of a contiguous decade which began when scoring was at its highest level since 1930, more than 10 percent higher than it is today.
Pitches per plate appearance have risen slowly but steadily over the past two decades, spending more than half that time in the 3.7-3.8 range. While the overall percentage of pitches which were strikes has barely moved, what's interesting is that number of balls per PA has increased at only about half the rate of the number of strikes per PA over the longer span (4.0 percent compared to 7.6 percent). Over the shorter span, the number of balls has actually decreased, which does jibe with what we'd expect as offense has deflated. In fact, of all of these various rates, B/PA is the only one which has a substantial correlation with scoring rates , r = .53, compared to .12 for S/PA, .15 for looking strikes (L/Str) and -.33 for in-play strikes (I/Str).
Breaking this down further, while the rates of swinging strikes (14-15 percent) and foul strikes (27 percent) have varied so little they're not worth including in the table above, the rate of looking strikes has risen by over 20 percent, to an all-time high ("all time" in this particular instance meaning "since 1988"). Correspondingly, the rate of in-play strikes has decreased by about 12 percent.
What's behind all of this? My theory is that two forces are in play here. First, Major League Baseball has spent the past decade attempting to enforce the official definition of the strike zone, directing umpires to call pitches strikes in a higher, narrower zone than they had been doing before. In 2001, MLB even introduced the controversial QuesTec pitch-tracking system in order to grade umps on their compliance, quietly replacing that with a newer and more widespread system called Zone Evaluation in 2009. According to a New York Times article about the latter, the Elias Sports Bureau's analysis of 2008 data concluded that umps in QuesTec parks called strikes slightly more often-a few pitches per game-than those not in such parks. Second, we've seen an evolution in the way strikeouts are understood within the game, at least beyond the Big Apple sports pages and their concern for David Wright. Thanks to statheads inside and out of front offices, we know that hitter strikeout rates correlate with a number of useful attributes, namely hitting for power and drawing walks. Furthermore, the accounting shows that hitters' failures to move runners over via "productive outs" are cancelled out by the decreased possibility of double plays, making a whiff no worse than any other kind of batting out.
Consider for a moment the history of Bobby Bonds' single-season record of 189 Ks, set in 1970. Players such as Three True Outcomes patron saint Rob Deer, Jose Hernandez and Preston Wilson gave chase to Bonds' record, but such was the stigmathat their managers sat them down late in the year to avoid breaking the record. Deer played in just five of the Brewers' final 12 games in 1987, settling for the AL record of 186. Hernandez played in just one of the Brewers' final four games in 2001, finishing at 185, and more dubiously sat for all but three of his team's final 11 games the following year, including games with playoff implications, while stuck at 188. Wilson had 185 Ks with six games remaining in 2000, but his pace was slowed by pinch-hitting in two games and missing the final game of the year to finish with 187.
Since all of that silliness, Bonds' record has been topped seven times by four men, Adam Dunn, Jack Cust, Ryan Howard and Mark Reynolds. Dunn broke Bonds' record in 2004, finishing with 195. Howard topped it with 199 in 2007, and Mark Reynolds blew through the 200-strikeout barrier in 2008 and 2009. This Magnificent Seven of Swish averaged 41 homers and 94 walks during those years, and none of the players who surpassed Bonds was reduced to parading around in public wearing a scarlet letter K.
While hitter strikeouts have become less stigmatized, pitcher strikeouts are valued ever more highly. For pitchers, missing bats means avoiding the largely inescapable vagaries of balls in play, a key element in preventing runs; if a hitter doesn't make contact, he can't even reach on an error, let alone collect a base hit or a homer. Furthermore, higher strikeout rates are tied to pitcher longevity and long-term success.
Recent work by researchers such as Tom Tango and J.C. Bradbury has shown that as far as strikeout rates go, hurlers tend to peak in their early 20s, somewhere between 20 and 25 years old according to Tango, depending upon the set of assumptions regarding regression which are used. A younger population of pitchers, it stands to reason, would generate higher strikeout rates, and as it turns out, the population of pitchers has indeed gotten ever so slightly younger in recent years.
From 1990 through 1997, the average pitcher age-July 1 age, weighted by innings pitched that season-hovered in the range between 28.1 and 28.5 years old. From 1998 through 2006 it fluctuated in the 28.5 to 28.9 range. It's dipped back below 28.5 for the past four years, the period when strikeout rates have risen most rapidly. The correlation between the average age and the fluctuating strikeout rates from 1990-2010 turns out to be almost nil (r = .01), though if we ditch the pre-expansion, pre-strike, pre-juiced ball years and isolate to 2000-2009, it shoots up (r= -.78; strikeout rates are rising while age is falling, remember). The youth movement is even more apparent if we limit the selection to ERA qualifiers (1 inning pitched per team scheduled game), as the average pitcher age fell nearly two full years from 2005 (29.3) to 2009 (27.6). It's back up this year, but hey, Stephen Strasburg is just getting started, and there are likely more minor league promotions to follow.
Where this becomes somewhat interesting, if not exactly a model of scientific rigor, is on the leaderboards. The average age of qualifying pitchers with above-average strikeout rates-say, 7.5 percent higher than league average (ignoring the relatively minor effects of park adjustment)-is dropping, and at a rate considerably faster than both the overall population of pitchers and the qualifiers:
Granted, we're only talking about 20 to 30 high-strikeout pitchers per year, and ignoring park effects, but the charts above suggests we're seeing a larger-than-normal influx of young high-strikeout pitchers, a class who by and large we can expect to be successful. Consider that the majors' current top 30 in strikeout rate (SO/PA) contains 11 pitchers 25 and younger, just seven 30 and over, and just one older than 33. Dial back to 2005 and you've got nine pitchers 25 and younger, 10 who were 30 and over, with four of those over 33 and two of them (Randy Johnson and Roger Clemens) over 40. Dial back to 2000, and it's 10 in the 25-and-under bucket, but 15 in the 30 and over one. The trend is probably more cyclical than linear, but it's certainly visible. Not all of those 11 youngsters (Clayton Kershaw, Mat Latos, Tommy Hanson, Yovani Gallardo, Phil Hughes, Felix Hernandez, Brandon Morrow, Ricky Romero, Max Scherzer, Ian Kennedy, Chad Billingsley) in this year's high-strikeout group are household names, but some of them will be someday, grabbing the headlines the way the 26-year-olds in that group (Ubaldo Jimenez, Josh Johnson, Tim Lincecum, Jon Lester, Francisco Liriano) are today.
My guess is that this is where a good bit of the talk about the Year of the Pitcher comes from. An especially large cohort of quality young hurlers who weren't on center stage a few years ago-not just high-strikeout pitchers 25 and under, mind you-is tasting success at a time when scoring is returning to levels not seen in almost two decades and strikeout rates are at an all-time high. Other changes in the game may be contributing to both trends, but they're the faces of the phenomenon, this so-called Year of the Pitcher.