Ruth, Bonds, Gehrig… Owings?

It’s no secret that Micah Owings is a great hitting pitcher, often causing analysts to refer to him as a hitter that happens to pitch rather than the aforementioned moniker. Reports even circulated prior to the season starting that Owings might get some playing time at first base due to the departure of Tony Clark. Last year, Owings produced one of the best hitting-seasons-for-a-pitcher of all time, thanks to a slash line of .333/.349/.683; he also hit eight doubles, one triple, and four home runs.

In the sixth inning of last night’s Diamondbacks-Astros game, Owings hit a pinch hit, two-run homer to tie the game. The Astros even made a pitching change prior to the at-bat in order to bring in righty Dave Borkowski and Brad Ausmus commented that Owings is the only pitcher over which he has ever discussed sequencing strategies. ESPN had a field day showing highlights and questioning whether or not Owings belongs in the lineup everyday, but the following video segment made me cringe:

They specially selected the ridiculously small sample size of 75 plate appearances in order to further a point that did not necessarily need to be made. Everyone knows he is a tremendous hitter and this comparison did nothing but show a complete ignorance towards the usage of statistics. The hard part about criticizing the video is that the anchors actually used and explained OPS! Granted, OPS is not the end-all, be-all, but for a mainstream show such as Sportscenter to discuss a sabermetric statistic is a pretty big step. Unfortunately, they lost points with the small sample size comparison.

Earlier today on PTI, Michael Wilbon mentioned that putting Owings in the lineup should be done sparingly at first until a large enough sample could be gathered to determine his true ability. Suffice it to say, I was shocked: One ESPN show discussed OPS and another discussed how small sample sizes should not be used to make quick judgments. While discussing sabermetric statistics and explaining how small sample sizes fail to explain anything truly tangible are both important, which do you feel would be best served exploring deeper on mainstream analysis-driven shows?

Print This Post

Eric is an accountant and statistical analyst from Philadelphia. He also covers the Phillies at Phillies Nation and can be found here on Twitter.

3 Responses to “Ruth, Bonds, Gehrig… Owings?”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. dan says:

    Definitely the small sample size issue. I actually got into an argument today (before walking away flabbergasted) about Phil Hughes’ ability. The other kid was saying how he’s an awful pitching prospect, and literally went down the Yahoo fantasy rankings pointing out every pitcher who was having a better season so far. By the way, he was totally serious. This person obviously watches ESPN, so maybe if ESPN occasionally talked about small sample sizes and how you can’t gain any information from them, he wouldn’t have sounded so ridiculous.

    Actually he’s a met fan, so there’s still something obviously wrong with him.

    Vote -1 Vote +1

  2. jlc says:

    I have to agree with the small sample sizes. It’s a human trait to see patterns that don’t really exist and to remember outliers as being more common than they are. You can’t really have any feel for statistics unless you understand the sample size issue.

    It’s not just baseball. It’s also things like the continuous splash of medical studies in the mainstream media that contradict each other, because of the small number of people involved in the studies. If everyone understood small sample sizes, a lot of life would be easier to navigate.

    Vote -1 Vote +1

  3. Eric Seidman says:
    FanGraphs Supporting Member

    Haha, nine out of ten doctors recommend it comes to mind with that. Well, how many groups of ten? Is it just 10? Is it 20? Is it 10,000 and 9,000 approve? I watched another clip I made from Redlasso and, on the 5AM Sportscenter, John Anderson basically eluded to the Owings graphic having a very small sample size. I wonder if someone got their attention and suggested a clarification.

    Vote -1 Vote +1