I have to think the average reader on this site does not fall prey to this kind of bias very often, although I don’t disagree that it’s something to be aware of, especially when you hear baseball players discussed in the media. In fact, I think I’m especially well aware of how much Delgado has turned it around in the past several months because of a slow start.
Using that example, I think we tend to forget that Delgado’s season line, which is certainly impressive, is now back in line with what we would expect from him. Because everyone left him for dead in April and May, it now seems remarkable that he’s putting up an .840 OPS on the year. However, that’s about where he was likely projected (I’m guessing) for this year.
Meanwhile, because Ellsbury had such a fast start, it’s not that baseball fans assume he is still doing well. Instead, we built up high expectations for his ’08 and as such are more liable to be disappointed with the results, quickly forgetting that this is his first full year in the majors and struggles were inevitable.
I think you make a valid point in this article, but you argue it from the wrong direction.
I think its almost impossible not to fall victim to small sample sizes at times because almost every media outlet in some form or fashion pushes them on you. Whether its team streaks or what a player is batting over the last week or series or vs this pitcher, its hard to block all of that out, knowing that it’s highly arbitrary and selective and ultimately doesn’t accurately portray the real picture.
This easily can be applied to fantasy baseball, when you are deciding which of two players to play for a particular game or week etc., you often use small meaningless sample sizes to convince yourself that one is the better option then the other, when in reality you can find small sample size numbers to support choosing either player and it simply comes downs to which one you “wanted” to play the entire time.
All this being said, if you do your best to avoid the ups and downs of variance within a season and just look at the final outcomes, you’ll be suprised how many were comfortably in the range of what you would have predicted before the season began, yet if you evaluate these same predictions during the season you will have a tendency to doubt yourself and your judgments based on these trvivial sample sizes that are merely traps setup to fool your mind.
Comment by Bill Krevski — September 8, 2008 @ 12:15 pm