I sent that e-mail about two weeks ago to my colleagues at Baseball Prospectus, and we had a discussion about whether what I was seeing -- I watch way too much baseball -- was real, whether it was a blip, and what the reasons for it might be. The general consensus was that while the numbers were down, it was too early to draw conclusions as to why.
With an additional couple of weeks of play in the books, and talk of a decline in offense making its way into the media, it's time to take a look and see what's actually happening. The lovely and talented Bil Burke -- well, the talented Bil Burke -- put together a significant amount of data and analysis for this piece, going beyond the basic stats we use and digging into the balls-in-play information we have for the past few years to look for answers (see chart, right).
The most significant conclusion from all this work is this: Through May 10 there is a massive disparity between the leagues in terms of the direction of change as compared to 2007. Offense is up in the NL, and way down in the AL. By runs per game, offense is down nearly 13 percent over two seasons in the American League, which is a staggering figure.
That the NL is outscoring the AL is a very strange phenomenon. The presence of the designated hitter has generally been worth about a half-run per game between the leagues, and while the gap has bounced around over the years -- closing at times due to changes in ballparks, talent distributions and random fluctuation -- for the NL to be outscoring the AL is quite odd. The Sporting News' David Pintodraws a connection between the gap and a bevy of young hitting talent in the NL. I'm not entirely convinced, but he lays out an interesting case.
Regardless, most of this study will focus on the AL's drop in run scoring. The rise in offense in the senior circuit is consistent with the variation we see from year to year in that league. The AL's figure is highly unusual -- the AL hasn't seen run levels like this in 16 seasons. What's down in the AL?
You should immediately notice the drop in slugging. From 2007 to '08 batting average has dropped by one point, but slugging has dropped by 15. This is in an environment with no new ballparks, and no radical changes to the existing ones. Over two years slugging is off by 40 points, isolated power by 28 points. At least OBP is moving in line with changes to batting average, which indicates that walk rates aren't the culprit for this cold spell at the plate. They've actually risen slightly, from 8.1 percent of PA in 2005 to 8.9 percent this year.
The availability of balls-in-play data over the last four years allows us to go a bit deeper. Think for a second about the things that go into run scoring. There's contact rate, or the amount of time a hitter doesn't strike out; batting average on balls in play, which is how often a struck ball becomes a hit; and both home run rate -- how often a struck ball leaves the yard -- and its older brother, extra-base hit rate. There's walk rate, which contributes to OBP. One measure that can be revealing is the number of extra-base hits as compared to the number of fly balls. Not all XBH come on fly balls, but most do.
Let's put these metrics and some others into a table to see where the changes are happening:
Batters are not striking out any more than they did last year, and only a bit more than they did in 2005 and '06. When they put the ball in play they get a hit about as often as they did last year, down slightly from '06. The drop from '07 to '08 has not been caused by some radical improvement in defending batted balls. Batters are improving slightly at controlling the strike zone, but it's a small change -- and it's not what we would expect given the lowered offensive environment.
What is driving the drop in runs? Well, fly-ball rates have dropped to a four-year low, and when batters do hit the ball in the air, they aren't getting as much bang for their buck. The rate of HR/FB is at a four-year low, and XBH/FB a three-year low; the drop from '07 to '08 ranges from five to 10 percent in both cases. American League batters are hitting fewer fly balls, and when they hit them, they're not getting as much production out of them. That's why slugging and home-run rates are way down, and with them, run scoring.
For what it's worth, while fly-ball rates are also down in the NL, the rates of HR/FB and XBH/FB are both up over 2007. National League batters are also hitting fewer fly balls, but their results are slightly better than they were last year, and in line with the previous two years. As in the AL, K/BB is down slightly, as is strikeout rate.
It has to be said that the change in run scoring and fly-ball outcomes could just be a blip. I haven't controlled for schedules or weather, and while I have not perceived a notable cold snap this spring, it is possible that balls are not flying as far because the conditions have not allowed for power. It could also be that we've had a distribution of games, more in pitchers' parks, that has skewed the numbers. Right now, all of this information falls into the category of "interesting, let's see more."
Speculation inside the game, as Buster Olney referenced in his blog on Saturday, centers on the theoretical eradication of performance-enhancing drugs from the game in the wake of the Mitchell Report. Personally I dismiss this out of hand. Since 2003, when survey testing kicked off a series of regimes and punishment mechanisms, offensive levels, and specifically power, have jumped around from year to year independent of what rules were in place. There has been no correlation between increased testing and greater penalties, and offensive levels and power, over the five-year period.
This makes sense when you think about it. Both from the players who have been suspended for failing tests and the ones named in the Mitchell Report, we learned that PED use was not something confined to power hitters, nor even hitters. Even if PED use has been affected by the rules changes, there's little reason to think that it would show itself in lowered offense.
The biggest reason to dismiss this claim, though, is the league split. Runs per game, slugging and XBH/FB are all up in the National League, whose players are subject to the same testing program as the ones in the AL. To assert that the overall falloff is due to the Mitchell Report and the impact of PED testing is to imply that all the juicers were American League hitters. That doesn't strain credulity; it causes credulity to laugh at you, smack you upside the head, and go find your best friend to smack him, too. The one explanation that I can safely rule out is some kind of Mitchell Report/testing effect.
So...why is power down? The drop in fly-ball rates combined with the drop in strikeout rates is an interesting phenomenon. We hear a lot about "pitching to contact," and you might speculate that the industry is shifting to a pitching model that emphasizes getting ahead in the count, keeping the ball down and in play, and getting better defense on the field to support that approach. Now, BABIP hasn't changed all that much, but fly-ball rates have. Because the batting average on fly balls in play is generally lower than the BA on groundballs in play, a constant BABIP and a lower fly ball rate implies that defense has gotten better. The lowered rate of extra-base hits per fly ball could imply improved outfield defense as well. Let's tease out double and triples and see what we get (see chart, right).
Meh. The rates are down compared to last year, but up over 2005 and '06. It is interesting to note that doubles and triples as a percentage of XBH (or if you prefer, doubles and triples relative to homers) have been rising since 2006. This is another indication that batters aren't getting as much bang for their buck on their fly balls.
Let's continue to speculate. If pitchers are aware that their mistakes will not go as far, and in fact, are much more likely to stay in the yard, they can worry less about contact. If pitchers are throwing more hittable pitches, strikeouts would decrease relative to batted balls, which is what we've seen. Walk rates are up very slightly in both leagues, which isn't the direction you'd expect in a lowered-strikeout/higher-contact environment, but the number is so small it could just be noise. The increase in groundballs may be a blip, but it's such a big jump that it seems like more, whether a pitching approach or a hitting approach, I do not know.
I think there may be some selection effects happening here. I've written about this when it comes to playoff baseball. Playoff games are generally lower scoring than regular-season games. However, part of the reason for that is that managers play as if they will be lower scoring, using more one-run strategies than they normally would and emphasizing defense to a greater extent. It becomes, if not a self-fulfilling prophecy, one that gets helped along.
Over the last year or so we've heard a lot about teams getting away from the style of baseball played during the peak of the high-offense era, and trying to play better defense. Personnel decisions along the lines of playing Tony Pena Jr. or Asdrubal Cabrera add up, and they start to impact the league's statistics. Teams have been choosing defense over offense, and that is probably the biggest reason for the drop in offense in the AL: personnel selection. Managers and GMs are putting lesser hitters on the field in an effort to prevent runs, and they're getting just that result -- for themselves and the opposition.
It's worth mentioning the possibility of an externality here. Anytime numbers such as HR/FB or XBH/FB go a little haywire, the baseball is a potential issue. If the ball is deadened even a little, whether by chance or by choice, it would have an effect on those numbers above all others. The increase in groundballs being hit would not be affected, however, and with that being such an important part of the AL's downturn in offense, it's reasonable to say that a change in the baseballs is an unlikely cause for the drop in run scoring.
The fall-off in offense in 2008 is notable, and the most interesting thing about it is the league split. While run scoring and the underlying indicators are higher in the National League, within a normal range of variation for those figures, there's been a large drop in the AL, driven by a drop in the number of fly balls hit and what happens when they are hit. Trends within the game that have nothing to do with performance-enhancing drugs, such as approaches to pitching and the distribution of playing time, are the most likely factors in this decline.
We are just six weeks into the season, and because of that, all of these figures are prone to dramatic change, so take these conclusions with a grain of salt, something to season your thoughts with the next time you see a 2-1 game with nine singles and a triple.