Predicting walk rate?

Most Cardinals fans out there know how special of a player Colby Rasmus is. He hits for power, doesn’t strike out that much, shows very good baserunner ability, and plays excellent defense in center. So far this year, he has contributed 2.4 Wins Above Replacement according to FanGraphs, which ranks him amongst some the best center fielders in the game. Also, at just 22 years old, he still has plenty of room for improvement.

The one problem with Rasmus’ game this year has been his walk rate. Always a patient hitter in the minors, with a career walk rate of nearly 12%, he has only drawn a free pass in 6.1% of his plate appearances this year. However, if you watch him player, it seems like he should be walking more. He has a selective eye at the plate, and doesn’t seem to lunge at too many pitches. The stats back it up also. According to Fangraphs’ O-Swing, he has swung at just 23.9% of pitches out of the strike zone compared to a league average rate of 25.0%

So why is he walking at such a low rate? Well, it occurred to me that the answer could very well be luck. He’s only had 335 plate appearances this year, and it seems entirely possible that he may be doing the things he needs to do to walk, but is experiencing bad timing.

Well to test the validity of my theory, I went to Fangraphs and downloaded the stats for each player from 06-08 who had at least 400 at bats in an individual year. I then plotted their O-Swing% and BB%:

image

That is a .686 correlation, which is pretty strong. Basically, this means that about 69% of a players ability to walk is tied to his ability to lay off pitches outside the strike zone. Intuitively, that makes sense; however, consider some of the other factors that you would expect to go into walking:

  • Pitches seen in the strike zone
  • Contact ability
  • Total pitches swung at
  • But when I tested those factors (Zone%, Contact, Swing%), neither returned nearly as high of a correlation as O-Swing, with the closest being Swing%. However, It would seem that with the right combination of those factors, along with some others that Fangraphs offers, you could essentially predict walk rate for any hitter. That would particularly be useful in the situation that a prefaced my article with, in which a players’ walk rate might be influenced by luck in a small sample size.

    I’ll play enough around with the data a little bit and see if I can come up with anything. Maybe, a certain fellow THT writer, who just wrote a couple nice pieces on regression to the mean, could point me in the right direction…


    10 Comments
    Oldest
    Newest Most Voted
    Inline Feedbacks
    View all comments
    John Burnson
    14 years ago

    Might a curved trendline fit better? Don’t think that every regression has to be linear.

    Nick Steiner
    14 years ago

    An exponential curve yields a .73 correlation.

    Eric/OR
    14 years ago

    Cool line of research.

    Nick Steiner
    14 years ago

    @Luis

    If you go to any player’s stats on FanGraphs, you can click “show averages” on most of the headings.

    Eric Johnson
    14 years ago

    Don’t you think that hitting in the 2 hole with Pujols behind him could have something to do with his walk rate?

    Luis
    14 years ago

    You stated that the avg player swings at 25% of pitches OOZ. From where does this information come? Not nit picking, just curious as i would like to see it. Thanks. This is great stuff!

    studes
    14 years ago

    Ah, I missed that Nick.  Thanks.

    studes
    14 years ago

    It can be useful to define stats that not everyone is familiar with.  O Swing is percentage of balls outside the strike zone that are swung at, right?

    Nick
    14 years ago

    Studes

    Yeah, that’s the right definition.

    Nick
    14 years ago

    Although I did say, “According to Fangraphs’ O-Swing, he has swung at just 23.9% of pitches out of the strike zone compared to a league average rate of 25.0%”