Two Pitcher xBABIPS

Thanks to two excellent researchers, we have two different pitcher xBABIPs nestled within the posts on this site.

Matt Swartz has an xBABIP that’s part of SIERA. He was nice enough to re-run it with all pitchers with 40 innings pitched or more for 2003-12. He did not adjust for usage by weighting pitchers with high IP more heavily, but he did regress BABIP using factors that his research has shown to influence the number. His pitcher xBABIP formula is:

BABIP = .290 + .045*GB% – .103*K%.

Steve Staude used batted ball data to find a pitcher xBABIP in the Community Blog a while back. Of course, the weakness of a model using line drive percentage is the fickleness of that stat from stadium to stadium and year to year, but it does make a lot of sense intuitively: give up fewer frozen ropes and you’ll give up fewer hits. His regression led him to the following pitcher xBABIP formula:

xBABIP = 0.4*LD% – 0.6*FB%*IFFB% + 0.235

To help out the fantasy player looking at pitchers for 2013, I put all pitchers with more than 100 innings pitched on one leaderboard featuring their BABIP and both xBABIPs. I sorted by the difference between their Line Drive xBABIP and their SIERA xBABIP.

Enjoy!

Name GB% FB% IFFB% LD% BABIP SIERA BABIP LD BABIP DELTA
Chris Young 22.3% 58.2% 15.3% 19.5% 0.287 0.2833 0.2596 -0.0238
Ross Detwiler 50.8% 32.8% 9.6% 16.4% 0.263 0.2971 0.2817 -0.0154
Phil Hughes 32.4% 47.6% 15.6% 19.9% 0.286 0.2837 0.2700 -0.0136
Bartolo Colon 45.7% 36.3% 11.3% 18.0% 0.286 0.2958 0.2824 -0.0134
Marco Estrada 34.3% 45.4% 18.0% 20.3% 0.298 0.2793 0.2672 -0.0121
Derek Holland 43.1% 40.1% 10.5% 16.8% 0.261 0.2889 0.2769 -0.0120
Aaron Laffey 49.8% 31.2% 11.5% 18.9% 0.258 0.3009 0.2891 -0.0118
Trevor Cahill 61.2% 22.7% 9.2% 16.1% 0.289 0.2984 0.2869 -0.0115
Barry Zito 40.4% 39.6% 13.4% 20.0% 0.281 0.2935 0.2832 -0.0103
Bruce Chen 32.7% 44.8% 17.6% 22.4% 0.304 0.2873 0.2773 -0.0100
Hector Noesi 37.0% 45.3% 9.2% 17.8% 0.266 0.2912 0.2812 -0.0100
Jordan Lyles 53.9% 29.0% 8.8% 17.1% 0.301 0.2980 0.2881 -0.0099
Jeff Francis 50.3% 31.0% 12.0% 18.8% 0.341 0.2971 0.2879 -0.0092
Gavin Floyd 47.2% 34.7% 12.2% 18.1% 0.299 0.2907 0.2820 -0.0087
Henderson Alvarez 57.0% 24.3% 8.1% 18.7% 0.291 0.3056 0.2980 -0.0076
Colby Lewis 33.0% 45.8% 16.1% 21.2% 0.279 0.2824 0.2756 -0.0068
Jason Vargas 40.2% 40.5% 11.3% 19.4% 0.254 0.2917 0.2851 -0.0066
Tommy Hunter 45.4% 34.6% 12.0% 20.0% 0.298 0.2966 0.2901 -0.0065
Matt Moore 37.4% 42.9% 14.3% 19.6% 0.293 0.2830 0.2766 -0.0064
J.A. Happ 44.0% 38.9% 10.1% 17.1% 0.315 0.2861 0.2798 -0.0063
Travis Blackley 47.7% 34.7% 7.8% 17.5% 0.268 0.2950 0.2888 -0.0062
Kyle Kendrick 46.5% 35.8% 8.5% 17.7% 0.278 0.2932 0.2875 -0.0057
Ryan Vogelsong 43.5% 38.0% 11.1% 18.5% 0.284 0.2889 0.2837 -0.0052
Hiroki Kuroda 52.3% 29.5% 10.4% 18.2% 0.281 0.2943 0.2894 -0.0049
Blake Beavan 36.6% 41.7% 12.3% 21.8% 0.282 0.2957 0.2914 -0.0042
Cliff Lee 45.0% 36.9% 11.8% 18.1% 0.309 0.2851 0.2813 -0.0038
Jake Peavy 36.5% 44.6% 11.2% 18.9% 0.272 0.2838 0.2806 -0.0031
Clayton Richard 53.8% 27.8% 5.8% 18.4% 0.272 0.3021 0.2989 -0.0031
Aaron Harang 38.6% 41.0% 12.1% 20.5% 0.277 0.2902 0.2872 -0.0029
Brian Duensing 47.3% 32.7% 10.7% 20.0% 0.319 0.2962 0.2940 -0.0022
Matt Harrison 49.0% 30.9% 11.0% 20.1% 0.284 0.2964 0.2950 -0.0014
Chris Capuano 40.3% 39.1% 12.9% 20.5% 0.284 0.2877 0.2867 -0.0010
Johan Santana 33.1% 42.9% 19.3% 24.0% 0.301 0.2820 0.2813 -0.0007
Derek Lowe 59.2% 21.0% 5.5% 19.8% 0.326 0.3078 0.3073 -0.0005
Ervin Santana 43.2% 37.3% 9.7% 19.5% 0.241 0.2915 0.2913 -0.0002
Jerome Williams 53.6% 28.2% 6.6% 18.2% 0.293 0.2965 0.2966 0.0001
Brandon Morrow 41.1% 39.9% 10.1% 19.0% 0.252 0.2865 0.2868 0.0004
Josh Beckett 42.5% 36.7% 12.4% 20.8% 0.294 0.2905 0.2909 0.0004
Mat Latos 45.6% 36.1% 9.0% 18.4% 0.266 0.2883 0.2891 0.0008
Clayton Kershaw 46.9% 34.0% 12.2% 19.1% 0.262 0.2849 0.2865 0.0016
Kevin Millwood 44.7% 33.5% 12.9% 21.8% 0.304 0.2942 0.2963 0.0021
Dan Haren 39.6% 39.6% 11.5% 20.7% 0.302 0.2883 0.2905 0.0022
Luis Mendoza 52.1% 27.1% 10.6% 20.8% 0.31 0.2983 0.3010 0.0027
R.A. Dickey 46.1% 34.1% 12.7% 19.8% 0.275 0.2852 0.2882 0.0030
Kevin Correia 51.2% 29.1% 5.4% 19.6% 0.274 0.3005 0.3040 0.0035
Bronson Arroyo 41.4% 37.5% 10.2% 21.1% 0.286 0.2927 0.2965 0.0038
A.J. Burnett 56.9% 24.3% 8.5% 18.8% 0.294 0.2938 0.2978 0.0040
Jeremy Hellickson 41.8% 37.2% 10.4% 21.0% 0.261 0.2916 0.2958 0.0042
Kris Medlen 53.4% 28.2% 8.6% 18.5% 0.261 0.2902 0.2944 0.0042
Madison Bumgarner 47.9% 33.3% 8.7% 18.8% 0.276 0.2884 0.2928 0.0044
Justin Masterson 55.7% 25.0% 7.0% 19.3% 0.309 0.2969 0.3017 0.0048
Carlos Villanueva 36.7% 44.0% 9.3% 19.2% 0.275 0.2824 0.2872 0.0048
Tim Hudson 55.5% 25.2% 4.1% 19.3% 0.27 0.3010 0.3060 0.0050
Edwin Jackson 47.3% 35.8% 3.6% 16.8% 0.278 0.2893 0.2945 0.0051
Wei-Yin Chen 37.1% 42.1% 10.1% 20.8% 0.274 0.2873 0.2927 0.0054
Ricky Romero 53.5% 26.4% 7.2% 20.1% 0.311 0.2986 0.3040 0.0054
Jon Lester 49.2% 28.8% 14.4% 22.0% 0.312 0.2926 0.2981 0.0055
Alex Cobb 58.8% 21.2% 9.3% 20.0% 0.295 0.2973 0.3032 0.0059
Mike Minor 35.4% 43.7% 10.4% 20.9% 0.252 0.2854 0.2913 0.0059
Jake Westbrook 58.1% 21.0% 8.3% 20.8% 0.312 0.3016 0.3077 0.0061
Wandy Rodriguez 48.0% 31.6% 8.2% 20.5% 0.28 0.2952 0.3015 0.0062
Joe Saunders 43.1% 35.6% 9.2% 21.3% 0.305 0.2939 0.3005 0.0066
Ian Kennedy 37.3% 42.1% 10.0% 20.6% 0.306 0.2854 0.2921 0.0068
Lucas Harrell 57.2% 22.5% 8.2% 20.3% 0.289 0.2983 0.3051 0.0068
Carlos Zambrano 49.1% 28.3% 13.3% 22.6% 0.284 0.2955 0.3028 0.0073
Josh Tomlin 41.6% 37.2% 7.4% 21.2% 0.309 0.2959 0.3033 0.0073
Dillon Gee 50.3% 29.6% 9.6% 20.1% 0.301 0.2910 0.2984 0.0073
Matt Cain 37.4% 41.7% 10.8% 20.9% 0.259 0.2842 0.2916 0.0074
James Shields 52.3% 29.0% 7.5% 18.7% 0.292 0.2892 0.2968 0.0075
Zach McAllister 40.5% 40.3% 7.0% 19.2% 0.304 0.2873 0.2949 0.0076
Justin Verlander 42.3% 35.6% 15.4% 22.2% 0.273 0.2833 0.2909 0.0076
Matt Garza 47.3% 33.2% 8.7% 19.5% 0.271 0.2880 0.2957 0.0077
Homer Bailey 44.9% 35.4% 7.1% 19.7% 0.29 0.2904 0.2987 0.0083
League Average 45.1% 34.0% 10.0% 20.9% 0.297 0.2899 0.2982 0.0083
Jon Niese 48.3% 30.7% 10.3% 21.0% 0.272 0.2914 0.3000 0.0086
Jered Weaver 36.0% 42.8% 9.4% 21.1% 0.241 0.2864 0.2953 0.0088
C.J. Wilson 50.3% 29.9% 7.4% 19.9% 0.281 0.2920 0.3013 0.0093
Ricky Nolasco 46.6% 31.7% 8.6% 21.6% 0.309 0.2955 0.3050 0.0095
Clay Buchholz 47.6% 32.9% 4.2% 19.5% 0.283 0.2948 0.3047 0.0099
Miguel Gonzalez 34.9% 42.6% 10.5% 22.4% 0.26 0.2875 0.2978 0.0103
Jaime Garcia 53.7% 25.9% 7.2% 20.3% 0.339 0.2946 0.3050 0.0104
Chris Volstad 49.2% 28.1% 9.2% 22.7% 0.315 0.2998 0.3103 0.0105
Rick Porcello 53.2% 22.6% 15.8% 24.2% 0.344 0.2998 0.3104 0.0105
David Price 53.1% 27.0% 9.2% 19.9% 0.285 0.2887 0.2997 0.0110
Jose Quintana 47.2% 31.1% 7.5% 21.7% 0.299 0.2965 0.3078 0.0113
Anibal Sanchez 46.4% 32.1% 10.2% 21.5% 0.31 0.2899 0.3014 0.0115
Jeremy Guthrie 40.8% 36.1% 9.5% 23.1% 0.294 0.2952 0.3068 0.0116
Johnny Cueto 48.9% 29.4% 10.1% 21.7% 0.296 0.2923 0.3040 0.0117
Scott Diamond 53.4% 25.6% 4.0% 21.0% 0.292 0.3011 0.3129 0.0118
Freddy Garcia 40.2% 34.8% 16.7% 25.0% 0.297 0.2882 0.3001 0.0119
Cole Hamels 43.4% 35.1% 11.9% 21.5% 0.29 0.2839 0.2959 0.0121
Paul Maholm 51.2% 27.5% 7.7% 21.3% 0.281 0.2947 0.3075 0.0128
Tommy Hanson 39.8% 39.4% 8.0% 20.7% 0.314 0.2861 0.2989 0.0128
Travis Wood 34.3% 43.9% 8.6% 21.8% 0.244 0.2866 0.2995 0.0130
Bud Norris 39.2% 39.6% 9.4% 21.2% 0.301 0.2845 0.2975 0.0130
CC Sabathia 48.2% 30.7% 10.2% 21.1% 0.288 0.2873 0.3006 0.0133
Jason Marquis 52.5% 26.7% 4.5% 20.7% 0.307 0.2969 0.3106 0.0137
Hisashi Iwakuma 52.2% 27.3% 6.0% 20.5% 0.282 0.2934 0.3072 0.0138
Jason Hammel 53.2% 28.1% 3.2% 18.7% 0.291 0.2904 0.3044 0.0141
Luke Hochevar 43.3% 35.0% 8.0% 21.7% 0.315 0.2909 0.3050 0.0141
Joe Kelly 51.7% 27.5% 4.4% 20.8% 0.306 0.2964 0.3109 0.0146
Doug Fister 51.0% 26.7% 10.9% 22.3% 0.296 0.2919 0.3067 0.0148
Ryan Dempster 43.5% 35.7% 7.3% 20.8% 0.277 0.2876 0.3026 0.0149
Edinson Volquez 50.6% 28.2% 8.5% 21.2% 0.292 0.2904 0.3054 0.0150
Mark Buehrle 41.3% 36.3% 7.4% 22.4% 0.27 0.2930 0.3085 0.0155
Tommy Milone 38.1% 37.3% 12.9% 24.7% 0.31 0.2893 0.3049 0.0156
Nathan Eovaldi 45.5% 31.3% 8.9% 23.2% 0.317 0.2952 0.3111 0.0159
Roy Halladay 44.7% 32.3% 11.3% 23.0% 0.301 0.2891 0.3051 0.0160
Randy Wolf 42.9% 34.2% 7.9% 22.9% 0.337 0.2940 0.3104 0.0164
Philip Humber 34.9% 42.9% 7.9% 22.2% 0.294 0.2868 0.3035 0.0167
Zack Greinke 49.2% 29.1% 9.1% 21.7% 0.306 0.2885 0.3059 0.0175
Yu Darvish 46.2% 31.6% 12.3% 22.2% 0.295 0.2829 0.3005 0.0176
Ubaldo Jimenez 38.4% 38.2% 9.5% 23.4% 0.309 0.2889 0.3068 0.0179
Kyle Lohse 40.5% 35.6% 9.9% 23.9% 0.262 0.2911 0.3095 0.0183
Brandon McCarthy 40.5% 35.1% 10.5% 24.4% 0.295 0.2922 0.3105 0.0183
Jordan Zimmermann 43.4% 33.4% 9.7% 23.2% 0.288 0.2900 0.3084 0.0184
Chris Sale 44.9% 32.0% 12.2% 23.0% 0.294 0.2846 0.3036 0.0190
Chad Billingsley 45.4% 33.1% 6.1% 21.5% 0.308 0.2896 0.3089 0.0193
Shaun Marcum 35.4% 41.5% 9.3% 23.1% 0.28 0.2846 0.3042 0.0196
Felix Hernandez 48.9% 28.6% 10.4% 22.5% 0.308 0.2875 0.3072 0.0197
Scott Feldman 42.2% 31.9% 14.7% 25.9% 0.318 0.2906 0.3105 0.0199
Jake Arrieta 43.8% 32.4% 11.8% 23.8% 0.32 0.2871 0.3073 0.0202
Ivan Nova 45.2% 32.4% 7.7% 22.4% 0.331 0.2892 0.3096 0.0204
Max Scherzer 36.5% 41.5% 10.6% 22.1% 0.333 0.2761 0.2970 0.0209
Gio Gonzalez 48.2% 30.0% 8.4% 21.9% 0.267 0.2857 0.3075 0.0217
Vance Worley 46.0% 29.8% 9.8% 24.2% 0.34 0.2921 0.3143 0.0222
Patrick Corbin 45.7% 31.0% 7.7% 23.3% 0.317 0.2911 0.3139 0.0228
Wade Miley 43.3% 33.7% 6.4% 23.0% 0.293 0.2912 0.3141 0.0229
Mike Leake 48.9% 26.6% 8.3% 24.5% 0.306 0.2962 0.3198 0.0235
Yovani Gallardo 47.7% 31.5% 4.0% 20.9% 0.29 0.2871 0.3110 0.0240
Francisco Liriano 43.8% 34.8% 5.4% 21.3% 0.3 0.2849 0.3089 0.0240
James McDonald 39.2% 39.4% 4.3% 21.4% 0.269 0.2858 0.3104 0.0246
Joe Blanton 44.6% 32.0% 7.4% 23.4% 0.31 0.2889 0.3144 0.0255
Felix Doubront 43.7% 33.0% 8.6% 23.4% 0.312 0.2854 0.3116 0.0262
Adam Wainwright 50.8% 26.3% 6.6% 23.0% 0.315 0.2901 0.3166 0.0265
Josh Johnson 46.2% 30.2% 7.2% 23.6% 0.302 0.2895 0.3164 0.0269
Jarrod Parker 44.3% 30.1% 10.6% 25.6% 0.29 0.2908 0.3183 0.0275
Erik Bedard 43.3% 33.2% 6.6% 23.4% 0.314 0.2876 0.3155 0.0278
Stephen Strasburg 44.2% 33.1% 9.2% 22.7% 0.311 0.2788 0.3075 0.0287
Lance Lynn 43.8% 32.2% 8.4% 24.0% 0.321 0.2848 0.3148 0.0300
Jeff Samardzija 44.6% 33.1% 3.8% 22.3% 0.296 0.2844 0.3167 0.0322
Tim Lincecum 45.9% 30.3% 3.8% 23.8% 0.309 0.2870 0.3233 0.0363
Mike Fiers 32.7% 39.2% 10.8% 28.2% 0.319 0.2789 0.3224 0.0435

You can download the full spreadsheet here if you like.




Print This Post

Graphs: Baseball, Roto, Beer, brats (OK, no graphs for that...yet), repeat. Follow him on Twitter @enosarris.


9 Responses to “Two Pitcher xBABIPS”

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

    I love this. Thanks.

    I play in a simulation league, where I suspect batted ball data are some of the main factors on how pitchers perform. I plan to use this to target pitchers who are likely to match or out-perform their conventional stats in simulation by comparing their LD BABIP to their actual BABIP.

    My own attempt at finding xBABIP was (((GB-IFH) * SeasonalConstant1) + ((FB-HR-IFB) * SeasonalConstant2) + (LD * SeasonConstant3) + IFH +BUH) / (GB+FB+LD+BU-BUH)
    The constants I grabbed from somewhere here are the hit rates: .159 for GB-IFH, .121 for OFFB, and .75 for LD (for 2012).

    A sample:
    Chris Young .2497
    Ross Detwiler .2693
    Phil Hughes .2819

    Disclaimer: My file was created just before the end of the season and I can’t find where I grabbed the constants or the formula.

    Vote -1 Vote +1

  2. Steve Staude. says:

    Cool, Eno, thanks for this. I have a second article waiting to be published that talks about the effects of IP on xBABIP’s accuracy, and a third in the works that has formulas more oriented towards future BABIP predictions. I just tweaked the second article to compare my formula and SIERA’s.

    Regarding the fickleness of LD% from stadium to stadium — does anybody have an explanation of the causes of the differences in LD% from stadium to stadium? Colorado makes sense as to why it should be the highest, but why are Seattle and San Diego low but San Francisco high? I’m wondering if this has more to do with the interpretations of the individuals who are recording the batted ball types.
    http://www.fangraphs.com/guts.aspx?type=pf&season=2011&teamid=0

    Rally, I was tempted to use IFH as a factor in my formula, since it was obviously useful, but using it would have messed with what I was trying to accomplish (a defense-neutral BABIP estimator). If the goal is direct prediction of future BABIPs, though, it does seem like a good idea to consider IFH, assuming the pitcher’s defenders and fields are going to stay fairly consistent.

    Vote -1 Vote +1

    • Will H. says:

      Depending on how high or low in a stadium each team has the stringer judging batted ball types may affect LD% because of how different heights can skew the perception of whether something had a high or low arc. This is the article I remember, which in the comments section raises as many questions as answers but is suggestive of one possible reason: http://www.hardballtimes.com/main/article/when-is-a-fly-ball-a-line-drive/

      Vote -1 Vote +1

      • Steve Staude. says:

        Thanks Will, interesting article. It does make a lot of sense that the distinction between a line drive and a fly ball would be harder to make from a higher vantage point. It might even help explain why fly ball pitchers tend to be shown giving up more line drives than ground ball pitchers.

        Wyers’ reason for excluding the outliers makes sense but is very speculative, so it’s hard to put a lot of faith in the 0.38 r^2. I think there’s probably a lot more at work here, but not much of it has to do with the parks themselves (just the air density, the batter’s eye, and the weather make sense as factors, to me).

        I wonder what else influences the stringers, though — are they more likely to score a particular batted ball a line drive if it falls for a hit… or at least if an infielder attempts to catch it? Do they confuse hard hit liners that land further into the outfield with fly balls?

        Vote -1 Vote +1

  3. RMD says:

    Strasburg was one of the unluckiest pitchers in baseball last year. Uh oh.

    Vote -1 Vote +1

    • Steve Staude. says:

      Well, my formula didn’t think so. It pegged him at 0.308. He has a career 0.306, and it was 0.311 last year. There’s not a lot of innings pitched to judge by, though (251.1 career).

      Vote -1 Vote +1

  4. Drew says:

    For the second formula, could you plug in career rates for the batted-ball data and have a sort of absolute range of a player’s xBABIP? Assuming there’s no change in park or pitching philosophy.

    Vote -1 Vote +1

    • Steve Staude. says:

      Yeah, my formula was really derived from looking at career rates, anyway (well, since 2002, anyway, because that’s as far back as the batted ball info goes on FanGraphs). The article I’m waiting on to be published talks about the ranges of how accurate it is, depending on how many innings pitched you have to go by.

      Since popup rate, but especially LD%, are not very consistent from year to year, you’re better off using the average of years worth of data. Perhaps the more recent years should be weighted more heavily, though — I haven’t tested that, really.

      Vote -1 Vote +1

      • Drew says:

        Excellent stuff Steve, (and a wonderful article, Eno).

        I’m real excited to see what the analysis has to show. I’m just finally getting around to dipping my toes into the analytical side of the game, and I gotta say, the water’s fine.

        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>