from Amazon
Monday, July 12, 2010
Year of The Pitcher, 2010?
Albert Chen has this cover story article on SI:
There’s the across-the-board dip in offensive statistics: Through Sunday an average of 8.9 runs per game had been scored, down from the average of 9.3 through the same date last season. If that figure holds, this would be the first season with a per-game average under nine since 1992. Home runs are in similar decline—the rate of 1.85 per game would be the lowest since ‘93. Ditto for hits (the rate of 8.9 per team per game would be the lowest since ‘93) and overall ERA (4.16, on pace for the lowest mark since ‘92).
Comparing to the same date is good, so we see here that the drop is 4% based on half a season. Albert however does not give the same context for the other numbers he reports. Instead, he does the “on pace” thing, which I am totally against. As Phil notes:
Ignore the obvious problem that there will always be more pitchers “on pace” for a goal than actually reaching it (for instance, there are lots of players “on pace” for a 162-home run season after one game).
In any case, is that 0.20 runs per game per team (9.3 minus 8.9 divided by 2) that big a deal?
Albert and I had an email interview a few weeks ago, and the reason that I do only emails and never phone calls is so that everything I do and say is recorded, and so, I can repost everything I say to a member of the media. I realize that 1% of what I see gets published, so I get to post the other 99% here. What follows is everything I said to Albert on this issue:
From 1901 to 2009, which is 108 back-to-back seasons, there have been 36 times that the runs per game dropped by at least 0.13 runs per game. There have been 31 times that the runs per game increased by at least 0.13 runs per game. And another 41 times where the runs per game was within 0.13 runs per game. That we are currently witnessing a drop of 0.13 runs per game (with still 3 months to go, and still summer months to enjoy) is about as non-story as there is, other than the people’s desire to look for streak stories, or hot-and-cold stories.
I don’t know how many people are actually saying this [pitchers taking over, less power]. For the people who are saying this, then yes [they are overstating it]. For the people who are not saying this, they understand the situation. Now, you tell me whether the majority understand, or are overstating it. I presume the majority understand.
The drop is around the 33rd percentile (or 67th percentile, depending on your perspective). A story is when something is at the 95th or 99th percentile.
From 2004-2006, the runs per game went from 4.81 to 4.59 to 4.86.
From 2000-2002, it dropped by 0.52 runs in two seasons.
From 1987-1988, it dropped by 0.58 runs in one season, after previously increasing by 0.31 runs per game.
1976-79: 3.99 to 4.47 to 4.10 to 4.46. THOSE are big changes.
And the real “year of the pitcher” changes: 1962-1963: drop of 0.51 runs per game 1968-69: increase of 0.65 runs per game, followed by another 0.27 runs per game.
What is missing is historical perspective. And, that’s hard to believe, given the ease in which this data is available. We don’t need experts. We need less people to make stuff up.
This stuff happens in two seasons out of every three. It’s typical, not fluky, and certainly not a trend.
I don’t know a lot about Albert Chen’s other work, but his use of PITCHf/x data is scary. He knows enough to be dangerous. In both the Ubaldo Jimenez companion piece to the Year of the Pitcher piece, and in his Stephen Strasburg piece for SI a few weeks ago, he abused the meaning of the PITCHf/x data for those pitchers. He used “break” numbers from Gameday with complete disregard for the physics reality or consistency in what he was saying.
Maybe I should have added that sort of error to my PITCHf/x pitfalls article, but that was the first time I’d seen a prominent author make that error.