FanGraphs Baseball


RSS feed for comments on this post.

  1. Carson: what if you focus ONLY on the Dodger game in the pre-1966 time period? That data is apparently very solid, since we even have pitch data there. I’d like to see if there’s a difference.

    Comment by tangotiger — February 22, 2011 @ 5:21 pm

  2. I wasn’t aware of that. Where is it available?

    Comment by Carson Cistulli — February 22, 2011 @ 5:24 pm

  3. A shrinking of ballparks from 1950 to 1964 would help explain the data neatly. Looking at the last chart, the next thing I’d want to do is calculate the aggregate yearly HR/Batted Ball in all the parks that were replaced or added* at any time in these years versus in all those which weren’t. Comparing the two year-by-year will tell you if parks were shrinking.

    *Braves / County
    Sportsman’s / Memorial
    Shibe / Municipal
    Griifith I / II / RFK
    Ebbets / Coliseum / Dodger
    Polo Grounds / Seal / Candlestick
    + Wrigley (LA) / Dodger
    + Metropolitan
    + Colt
    + Polo Grounds / Shea

    Versus the other 10 franchises.

    Comment by Eric M. Van — February 22, 2011 @ 6:15 pm

  4. Oh, and thanks for the shout-out! This is fabulous work.

    Comment by Eric M. Van — February 22, 2011 @ 6:16 pm

  5. The “early spike” that you date 1950-64 actually seems to peak a bit later. Bear in mind that after 1968, when batting averages were very low (in the AL the top average was .301, the second-best .290) changes were made, notably to the mound, to increase offense. In other words, if the peak is about ’68, there could be a reason why the curve would rise to that point, then drop a bit and stabilize.

    Comment by Mr Punch — February 22, 2011 @ 6:21 pm

  6. Excellent work, thoroughly enjoyable and informative.

    Comment by Mark — February 22, 2011 @ 6:30 pm

  7. While it looks like xGB% does peak around 1969, it was on a clear and steady rise from the mid-1950s on.

    Considering the grouping over the past 8 years or so, my guess is that previous data simply wasn’t sufficient. Recent data points are grouped closer together than during any other time frame on the chart. More consistent data hints at (though doesn’t necessarily prove) more reliable data.

    Comment by Josh — February 22, 2011 @ 6:41 pm

  8. The rapid increase in the late 50s corresponds with a significant change in glove technology. Modern gloves with the edge-u-cated heel started replacing the old gloves that had been worn for generations in the late 1950s.

    I’m not sure how this change would affect GB rates, but it has to have some impact.

    Comment by EB in LA — February 22, 2011 @ 7:30 pm

  9. I’m humbled, sir — by your praise specifically, and by my shortcomings in attempting to wrestle this work to the ground. (And also, apparently, my shortcomings in constructing a decent metaphor with regard to my shortcomings.)

    As for the suggestion about ballparks, I’ll consider tackling it. But also: it looks hard. Would it be hard?

    Comment by Carson Cistulli — February 22, 2011 @ 7:39 pm

  10. A thought-provoking read.

    Is there any chance that integration plays a role in the style of baseball (rebirth of small-ball a la #42/Wills/Mays/Brock) or the impact of speed in the outfield in that period? My impression is that the expansion of the talent pool was weighted toward a huge influx of rangy outfielders. This coupled with the greater availability of hitters with plus speed might have fed into deliberate shifts in batting approaches or managerial strategy. The list of potential explanations for the difference in GO/AO over that span understates that it represents a rising trend, not just an era with a distinct mean. Option 4 mentions hitting approaches and the possibility that more flyballs were caught in ’50 than ’64. Perhaps the decreasing denominator in the ratio (AO) reflects that for a growing number of batters, it made sense to keep the ball on the ground.

    Comment by Oliver — February 22, 2011 @ 9:36 pm

  11. This excellent Rob Neyer article could help explain the increasing GB tendency. A major switch, as Neyer suggests happened in the 70’s, from curve to slider would result in a lot more ground balls, no? And it appears from the article that the pitch gained favor with pitchers over time.

    Comment by Paul — February 22, 2011 @ 9:38 pm

  12. I dunno if this actually made such a big difference, but this is an awesome insight.

    Comment by SeanP — February 22, 2011 @ 10:22 pm

  13. Retrosheet. What I mean to ask is if you repeat your xGB% chart (your first one), but ONLY with the Dodger games. Do you still get the same pattern or does it look smoother?

    Comment by tangotiger — February 23, 2011 @ 1:30 pm

  14. I wonder if pitch type usage matters. I know the fastball and change have been around forever, but when were other pitches invented? Perhaps as more pitch types were invented or perfected by some ground balls became more likely.

    Comment by Kick me in the GO NATS — February 23, 2011 @ 4:05 pm

  15. Whether it’s hard completely depends on the data set you’re working with, and / or how quick and comfortable you are processing Retrosheet data.

    It’s actually pretty easy, I think, and I might tackle it myself after the Academy Awards. I have eight movies to see before then (the curse of multiple obsessions strikes again!).

    Comment by Eric M. Van — February 23, 2011 @ 4:55 pm

  16. If you look at the data, ’50-’58 is one era, ’64 and onward is another, and ’59-’63 is the transition between them. If the use of the new gloves began significantly in ’59 and their use was universal by ’64, that would fit.

    It would make sense that improved glove technology would be much more important to infielders than outfielders, based on the frequency with which modern players get their glove on a ball but fail to make the play.

    Comment by Eric M. Van — February 23, 2011 @ 5:04 pm

  17. I think part of what you’re seeing may just be a change in the data.

    See the study “FIELDER ‘RANGE’ RATINGS” at

    and especially the graph entitled “Per Cent Frequency of Hit Descriptions”.

    Since the B-R data comes from Retrosheet I believe, this should be applicable.

    Comment by KJOK — February 24, 2011 @ 4:37 pm

  18. Looks like I’m late to the party, but is there a chance xGB% dating back to at least 1964 could be presented on the sight, and possibly something like XxFip (xFip with xGB%). That would be really awsome

    Comment by William — August 6, 2011 @ 7:09 pm

Leave a comment

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Close this window.

0.230 Powered by WordPress