And Here’s the Pitch …

This is a game to be savored, not gulped. There’s time to discuss everything between pitches or between innings.
-Bill Veeck

Last week The Washington Post printed a position-by-position analysis that detailed which players had swung and missed most and least frequently through August 20 of this season. This prompted a bit of discussion on the SABR-L listserv (and did I mention that if you’re not a member of SABR you should be?) as several members debated the values of this sort of data and the merit of analyzing it. Based on that discussion I thought this week I’d give a short synopsis of how pitch data is typically used and discuss what might and might not be learned from its analysis.

For this article I used Retrosheet data and included the 473 players who had at least 500 plate appearances in the years 2000 through 2004.

Pitches per Plate Appearance

Perhaps the most basic stat that can be derived from pitch data is the average number of pitches per plate appearance (P/PA). While total number of pitches is readily available on they do not calculate P/PA for you. In any case the leaders and trailers in pitches per plate appearance for 2000-2004 were:

Highest                 PA     OPS Pitches    P/PA
Rickey Henderson      1290     694    5595    4.34
Brad Wilkerson        2029     838    8740    4.31
Todd Zeile            2552     751   10953    4.29
Jeremy Giambi         1299     828    5574    4.29
Adam Dunn             2112     893    9039    4.28

Lowest                  PA     OPS Pitches    P/PA
Randall Simon         1425     733    4310    3.02
Brent Butler           597     664    1810    3.03
Nomar Garciaparra     2456     906    7638    3.11
Rey Sanchez           2196     627    6876    3.13
Deivi Cruz            2603     699    8197    3.15

Not surprisingly, the players in the first list have much higher on-base percentages (OBP) than those in the second since it’s difficult to take a walk when you’re swinging so much, a la the quintessential “bad-ball hitter” Randall Simon. However, that’s not to say that players who see fewer than 3.4 pitches per plate appearance (which includes about 8% of the players in this study) can’t be successful. Nomar Garciapara and Vladimir Guerrero at 3.28 P/PA are cases in point along with Garret Anderson (3.32 P/PA) and Vernon Wells (3.37 P/PA). These players are aggressive hitters who often make solid contact.

That said, the players in the bottom 20% of P/PA have an average on-base plus slugging (OPS) of 719 while those in the top 20% have an OPS of 797. Since OPS correlates very well with run production, it is therefore a good proxy, and because going deep into counts forces opposing pitchers to throw more pitches and tire sooner, it’s safe to say that on average, players who see more pitches end up contributing more to their teams.

Swinging at the First Pitch

A second aspect of pitch data that is often discussed concerns how often players swing at the first pitch. Here are the top and bottom five in this category:

Highest                 PA     OPS Pitches    P/PA    1stP     1stP/PA
Wily Mo Pena           563     771    2047    3.64     276       0.490
Nomar Garciaparra     2456     906    7638    3.11    1202       0.489
Karim Garcia           830     741    2813    3.39     403       0.486
Vinny Castilla        2738     734    8870    3.24    1306       0.477
Vladimir Guerrero     3168    1005   10399    3.28    1500       0.473

Lowest                  PA     OPS Pitches    P/PA    1stP     1stP/PA
Scott Hatteberg       2415     763    9828    4.07     229       0.095
Todd Zeile            2552     751   10953    4.29     263       0.103
Jason Kendall         3279     778   12842    3.92     341       0.104
Mark Ellis            1026     711    4167    4.06     121       0.118
Randy Velarde         1090     749    4358    4.00     130       0.119

Niether of these lists contains many surprises with Garciaparra and Guerrero making the top five and Scott Hatteberg and A’s teammates Jason Kendall and Mark Ellis in the bottom five.

What is most interesting is that the top 20% have an average OPS of 760 while the bottom 20% recorded an almost identical 759 OPS. Further, the correlation coefficient (the measure of the linear relationship of two measures) between first pitch percentage and OPS was 0.01. In other words, it seems that players adopt different strategies as to whether to swing at the first pitch but at the plate appearance threshold of this study it doesn’t make a difference in their overall productivity.

Swinging and Missing

The category the Post focused on was how often players swung and missed. Over the last five years the leaders and trailers in this category included:

Highest                 PA     OPS Pitches    P/PA    Miss      Miss/P
Wily Mo Pena           563     771    2047    3.64     400       0.195
Russ Branyan          1403     805    5774    4.12    1087       0.188
Jared Sandberg         706     703    2796    3.96     498       0.178
Todd Greene            726     734    2460    3.39     425       0.173
Ruben Rivera          1010     690    3840    3.80     646       0.168

Lowest                  PA     OPS Pitches    P/PA    Miss      Miss/P
Juan Pierre           3037     742   10160    3.35     287       0.028
Luis Castillo         3231     743   13015    4.03     389       0.030
David Eckstein        2520     700    9531    3.78     291       0.031
Chuck Knoblauch       1393     685    5484    3.94     172       0.031
Scott Hatteberg       2415     763    9828    4.07     316       0.032

Once again you see free swingers with high strikeout rates in the leaders and contact hitters in the trailers as you would expect. And because the free swingers tend to have higher slugging percentages and the contact hitters higher on base percentages, the tradeoff means there is little correlation between swinging and missing and OPS.

That correlation is slightly positive (.11) with the top 20% recording a 745 OPS while the bottom 20% a 770 OPS but once again we’re likely seeing players focusing on their strengths in order to succeed. Of course I wouldn’t be hitting and running with the players in the first list…

Fouling off Pitches

Next, we’ll consider the ability to foul off pitches. The leaders and trailers in foul balls per plate appearance are:

Highest                 PA     OPS Pitches    P/PA    Foul     Foul/PA
Kevin Young           1669     722    6699    4.01    1370       0.821
Johnny Estrada         899     741    2999    3.34     736       0.819
Joe McEwing           1169     644    4739    4.05     933       0.798
Vance Wilson           714     692    2709    3.79     559       0.783
Tomas Perez           1051     692    3947    3.76     815       0.775

Lowest                  PA     OPS Pitches    P/PA    Foul     Foul/PA
Dave Roberts          1316     690    5057    3.84     583       0.443
Bill Haselman          546     714    1916    3.51     244       0.447
Jason Tyner            844     594    2743    3.25     390       0.462
Mark McLemore         2119     712    8417    3.97     982       0.463
Tom Goodwin           1388     661    5364    3.86     644       0.464

While it is often said that being able to foul off pitcher’s pitches in order to extend an at bat is a skill a la Richie Ashburn or Ichiro Suzuki (who was in the middle of the pack at .643), over the long run it doesn’t appear that fouling off more pitches than average makes a player any more successful. The top 20% had an OPS of 767 while the bottom 20% were at 753 with the correlation even lower than that for swinging and missing.

The idea that fouling off pitches is a skill is likely related to the several instances over the last five years of players fouling off pitch after pitch only to finally succeed. The most memorable being the 18 pitch at bat in the bottom of the 7th that the Dodgers Alex Cora recorded against the Cubs Matt Clement on May 12, 2004 that resulted in a two-run homerun to right field. That pitch sequence went as follows:


were B=ball, C=called strike, F=foul, and X=put into play. For the record that’s fourteen consecutive foul balls.

Other long at bats that ended well for the batter include:

  • 15 pitches on 5/13/2000 ending in a homerun by Todd Helton off the Giants Mark Gardner
  • 16 pitches on 7/7/2004 ending in a solo homerun by Terrmel Sledge off Antonio Alfonseca
  • Plate Discipline

    The final category is the percentage of pitches taken for balls not including intentional balls. The leaders and trailers here include:

    Highest                 PA     OPS Pitches    P/PA    Ball      Ball/P
    Barry Bonds           3050    1316   12060    3.95    5377       0.446
    Chad Kreuter           634     757    2643    4.17    1166       0.441
    Rickey Henderson      1290     694    5595    4.34    2437       0.436
    Jason Giambi          3036    1020   12622    4.16    5468       0.433
    Mark McLemore         2119     712    8417    3.97    3645       0.433
    Lowest                  PA     OPS Pitches    P/PA    Ball      Ball/P
    Alex Gonzalez         2321     674    8289    3.57    2410       0.291
    Johnny Estrada         899     741    2999    3.34     880       0.293
    A.J. Pierzynski       2015     775    6430    3.19    1893       0.294
    Wilton Guerrero        649     629    2096    3.23     621       0.296
    Rod Barajas            924     645    3292    3.56     979       0.297

    Here is where we see the biggest difference with the top 20% recording an OPS of 837 and the bottom 20% at 705 and a correlation coefficient of .52.

    Naturally, you would expect this since OBP is half of OPS and the number of balls taken directly impacts OBP. In addition, the percentage of pitches taken for balls is not only a function of the selectivity of the batter but also the desire of the pitcher to not serve up a fat pitch to a good hitter as illustrated by the inclusion of both power hitters like Bonds and Giambi as well as the likes of Henderson and McLemore.

    The Verdict

    So is pitch data an important tool for performance analysis?

    My take is that it certainly can paint a picture of how a player approaches his at bats. I like to take a look at it, for example, when a player’s performance suddenly changes to see what they might be doing differently if anything. But as a tool for analysis these standard ways of looking at pitch data don’t really add much to the picture you get from aggregate statistics like OPS or Runs Created.

    Print Friendly
     Share on Facebook2Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
    « Previous: Not Just Any Bob Miller
    Next: Business of Baseball Report »

    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>