Real Groundball Rates

In his first start with the Angels last night, Joel Pineiro showed that the sinker he learned from Dave Duncan traveled with him to Anaheim, getting 15 groundballs in 6 innings of work. While he gave up three runs, he’ll pitch well for the Angels as long as that two-seam fastball is still diving, and there’s no indication that he lost it in the transition back to the AL.

However, this post isn’t really about Joel Pineiro. It’s about GB/FB ratio. Pineiro’s 15.00 GB/FB ratio tied with Felix Hernandez for the best of any pitcher to throw so far. However, if you sort by GB%, rather than GB/FB, you will notice that Ryan Dempster actually posted a higher groundball rate than Pineiro. Dempster ran a 75 percent GB% in his first start, compared to 71 percent for Pineiro, though his GB/FB ratio is a more pedestrian 4.50.

The difference, of course, is line drives. Dempster only gave up one, while Pineiro gave up five, and those don’t go in the GB/FB calculation. If the Twins had hit Pineiro less effectively, and those line drives had been converted into fly balls, his GB/FB ratio would have gone down, even though he would have pitched better and likely allowed fewer runs.

In other words, in this case, a higher GB/FB ratio is actually a bad thing.

Now, over a full season, this mostly evens out, and the correlation between GB% and GB/FB is pretty darn high. However, given that we have GB%, FB%, and LD%, I’d like to see GB/FB go by the wayside. It doesn’t serve any purpose, really. Why evaluate a pitcher on just two of the three batted ball types, rewarding him for giving up more of the most harmful kind?

GB% gives you all of the good information of GB/FB without any of the bad assumptions about line drive rates being equal. When describing a pitcher’s batted ball tendencies, you’re better off with GB% than GB/FB.




Print This Post



Dave is a co-founder of USSMariner.com and contributes to the Wall Street Journal.

8 Responses to “Real Groundball Rates”

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

    Even then, GB% assumes all FBs are equal. Wagner’s a good example, when he’s at his best his GB% is very low, but his IFFB/FB rate is pretty spectacular. Somebody needs to come up with a good number that composites GB/FB with LD% and IFFB rates.

    Vote -1 Vote +1

  2. tangotiger says:

    I prefer GB-FB per contacted ball. That basically treats LD as if they are half GB, half FB. So, these are all equivalent:

    GB FB LD
    60 30 10
    55 25 20
    50 20 30

    Vote -1 Vote +1

    • Robin says:

      With all due respect – why is it desirable to view a 60/30/10 split as equal to a 50/20/30 ?

      Aren’t those very different pitcher profiles (given sufficient sample size)?

      Aren’t LDs definitively the most damaging type of contact?

      Sincerely curious.

      Vote -1 Vote +1

      • tangotiger says:

        The question is purely on how to classify a pitcher as a GB or FB pitcher. There are no “LD” pitchers in MLB, as virtually all of them have a LD rate allowed of beween 17% and 23%. Basically, it’s impossible to sustain an LD rate higher than that (you’ll be out of baseball) or lower than that (hitters want to drive the ball to the gap, so they’ll adjust).

        Given that the LD rate settles at 20% for virtually all pitchers, the focus is therefore on the GB and FB. And if you have a pitcher that is 60/30/10, and we know his LD is going to settle in at 20%, then the question is: where are his GB and FB numbers going to settle?

        And I’m suggesting that a 60/30/10 pitcher will settle at 55/25/20. That is, he’ll maintain the GB-FB differential, rather than maintain the 60/30 ratio.

        Which is also Dave’s point here.

        Vote -1 Vote +1

  3. Joe says:

    crazy stuff going on with Rick Porcello’s pitch f/x today in Detroit, including a 97 mph fastball with 18″ of break and 42″ pfx…check it out

    Vote -1 Vote +1

  4. LibertyBoy says:

    At one point, the rating for pitchers that tells us how hard they are getting hit will be calculated by actually measuring the velocity and trajectory of batted balls. Until then, does anyone know how far on the horizon will any kind of qualitative tag be added to batted ball stats? I’ve seen some stuff on mlb.com’s gameday e.g. “sharp line drive”, “soft ground ball” but I haven’t seen that information work its way in the B-R’s of the world.

    Vote -1 Vote +1

  5. JMHawkins says:

    I think GB% is fine, since if you really want an overall measure of how hard a pitcher is getting hit, you might as well use xFIP, since not getting hit at all is a rather big part of how hard he’s being hit. GB% is good for analyzing a subset of a pitchers skillset, especially since for most mortals, a high K% means a low GB%. The rare guy who combines high K% with high GB% is special.

    I wonder about IFF%. Is that consistent year-to-year for guys? Or are the numbers just too small to get a clear picture. I’m going to need to look.

    Vote -1 Vote +1

  6. catbaby says:

    Instead of treating line drives as 50% GB and 50% FB, why not distribute them according to a pitcher’s GB/FB ratio?

    For example, let’s say a pitcher’s GB:FB is 2:1 gives up 10 more LD than he “should have”. Instead of putting 5 of the LDs as GBs and 5 of them as FBs, put 6.67 of them as GBs and 3.33 of them as FBs.

    So, if you have a pitcher who’s 60/30/10 but bound to settle at 20 LDs instead of 10, I say you award him 53.33/26.67/20.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *