Using Contact and Pull Rate to Predict a Batter’s Decline
My fellow RotoGraphs writers and I have been working through our 2016 player evaluations. I was tasked with the declining J.J. Hardy. I noticed that two of his core hitting traits, the ability to make contact and pull the ball, had degraded from 2013 to 2014. While Hardy saw an overall decline in his skills when these two skills declined, I wanted to see it was common among other players. I was able to find that decline in these two areas is a sign of a player on a fast decline.
I seeing pulling the ball and making contact as skills which show that a player has the ball speed and vision to keep hitting the ball. To put Hardy’s decline into perspective, here are his Contact% and Pull% values for the past four seasons to go along with his ISO and wRC+.
| Season | Pull% | Contact% | ISO | wRC+ |
| 2013 | 44.1% | 87.8% | 0.170 | 99 |
| 2014 | 41.4% | 81.5% | 0.104 | 90 |
| 2015 | 40.1% | 84.2% | 0.092 | 49 |
Both ISO and wRC+ dropped, but again this is the sample of one person.
To see if all players decline when contact and pull rate drop, I ran a simple study. I took the double the change in Contact% plus the change in Pull% to get an overall change value. Then I grouped the players together by changes of at least 10 percentage points to the good or bad. Going back to 2002, I looked to see how a combined decline affected the player’s stats and the chances for a future rebound (min 200 PA in season one and two). The normal average decline around age 29 is ~ 4 wRC+ and ~.003 ISO.
| Overall Value (2*Change in Contact%+Change in Pull%) | Age | wRC+ (Y1 to Y2) | wRC+ (Y2 to Y3) | Total Change in wRC+ | ISO (Y1 to Y2) | ISO (Y2 to Y3) | Total Change in ISO |
| Decline of 15% points or more | 29.3 | -11.4 | -7.1 | -18.5 | -0.011 | -0.014 | -0.025 |
| Decline of 10% points or more | 29.2 | -5.8 | -5.1 | -10.9 | -0.009 | -0.011 | -0.020 |
| Increase of 10% points for more | 29.0 | 3.0 | -8.6 | -5.7 | 0.003 | -0.009 | -0.007 |
| Increase of 15% points or more | 29.0 | 3.5 | -13.3 | -9.8 | -0.001 | -0.007 | -0.009 |
I will start off with the 15% Point increase group which doesn’t follow the pattern of the other three. I am not sure if this number is off because of a small sample size issue or if really cranking up the Pull% is the last gasp of a declining player. Otherwise, my inclinations of a declining player were correct. Once a player starts seeing a decline in Pull% and Contact%, their power and overall production decline more than average and there seems to be no regression/bounce back the next season.
With the above results in mind, here are the 2015 players who saw, at least, a 10% drop in the their overall value.
| NAME | Contact% Drop | Pull% Drop | Overall Value Change | 2014 wRC+ | 2015 wRC+ | wRC+ change | 2014 ISO | 2015 ISO | ISO change |
| Hunter Pence | -5.0% | -13.0% | -22.0% | 123.9 | 127.3 | 3.3 | 0.168 | 0.203 | 0.035 |
| Seth Smith | -7.0% | -8.0% | -21.0% | 132.5 | 113.4 | -19.0 | 0.174 | 0.195 | 0.021 |
| Sean Rodriguez | -2.0% | -17.0% | -21.0% | 97.6 | 78.6 | -19.0 | 0.232 | 0.116 | -0.116 |
| Miguel Montero | -8.0% | -2.0% | -19.0% | 90.0 | 107.0 | 16.9 | 0.127 | 0.161 | 0.035 |
| Victor Martinez | -5.0% | -9.0% | -18.0% | 167.2 | 77.2 | -90.0 | 0.230 | 0.120 | -0.109 |
| Matt Holliday | -4.0% | -8.0% | -17.0% | 132.0 | 124.3 | -7.7 | 0.169 | 0.131 | -0.038 |
| Brett Gardner | -5.0% | -5.0% | -16.0% | 111.2 | 105.2 | -6.0 | 0.166 | 0.140 | -0.026 |
| Travis Snider | -6.0% | -4.0% | -16.0% | 122.1 | 81.2 | -40.8 | 0.174 | 0.118 | -0.056 |
| Alex Avila | -2.0% | -12.0% | -16.0% | 97.4 | 82.5 | -14.9 | 0.141 | 0.096 | -0.046 |
| Rene Rivera | -4.0% | -7.0% | -15.0% | 113.4 | 33.0 | -80.4 | 0.180 | 0.097 | -0.083 |
| Adam LaRoche | -6.0% | -3.0% | -14.0% | 127.0 | 74.8 | -52.2 | 0.196 | 0.133 | -0.063 |
| Kole Calhoun | -6.0% | -1.0% | -13.0% | 124.6 | 104.7 | -19.9 | 0.178 | 0.167 | -0.012 |
| Edwin Encarnacion | -4.0% | -5.0% | -13.0% | 151.2 | 149.6 | -1.7 | 0.279 | 0.280 | 0.001 |
| Caleb Joseph | 0.0% | -11.0% | -12.0% | 71.0 | 86.3 | 15.3 | 0.146 | 0.159 | 0.013 |
| Jonathan Schoop | -3.0% | -6.0% | -12.0% | 64.1 | 111.5 | 47.4 | 0.145 | 0.203 | 0.058 |
| Brad Miller | -3.0% | -7.0% | -12.0% | 87.2 | 105.0 | 17.8 | 0.144 | 0.144 | -0.001 |
| Matt Carpenter | -10.0% | 7.0% | -12.0% | 117.1 | 138.9 | 21.8 | 0.103 | 0.233 | 0.131 |
| Brett Lawrie | -6.0% | 1.0% | -12.0% | 102.9 | 94.5 | -8.4 | 0.174 | 0.148 | -0.026 |
| Jean Segura | -2.0% | -9.0% | -12.0% | 66.1 | 62.3 | -3.8 | 0.080 | 0.079 | -0.001 |
| Casey McGehee | -6.0% | -1.0% | -12.0% | 102.2 | 52.4 | -49.8 | 0.070 | 0.076 | 0.006 |
| Chris Carter | -1.0% | -9.0% | -11.0% | 122.2 | 101.1 | -21.1 | 0.264 | 0.228 | -0.037 |
| Brock Holt | 0.0% | -11.0% | -11.0% | 97.6 | 98.2 | 0.6 | 0.100 | 0.099 | -0.001 |
| Ruben Tejada | -3.0% | -6.0% | -11.0% | 89.1 | 95.1 | 6.0 | 0.073 | 0.089 | 0.016 |
| Alcides Escobar | 0.0% | -11.0% | -11.0% | 94.3 | 66.6 | -27.7 | 0.092 | 0.064 | -0.028 |
| Andrew McCutchen | -3.0% | -5.0% | -11.0% | 169.0 | 146.4 | -22.6 | 0.228 | 0.196 | -0.032 |
| Nelson Cruz | -3.0% | -4.0% | -11.0% | 136.9 | 158.0 | 21.1 | 0.254 | 0.264 | 0.010 |
| Hank Conger | -5.0% | -2.0% | -11.0% | 82.1 | 107.4 | 25.3 | 0.104 | 0.219 | 0.115 |
| Adam Eaton | -5.0% | 0.0% | -10.0% | 117.1 | 118.4 | 1.3 | 0.101 | 0.144 | 0.043 |
| Eugenio Suarez | -1.0% | -8.0% | -10.0% | 86.0 | 104.8 | 18.8 | 0.094 | 0.167 | 0.072 |
| Jayson Werth | -5.0% | 0.0% | -10.0% | 140.9 | 87.2 | -53.7 | 0.163 | 0.163 | 0.000 |
| David Peralta | -1.0% | -8.0% | -10.0% | 110.1 | 138.1 | 28.0 | 0.164 | 0.210 | 0.046 |
| Ryan Zimmerman | -2.0% | -6.0% | -10.0% | 119.9 | 106.7 | -13.3 | 0.168 | 0.217 | 0.049 |
| Omar Infante | -5.0% | 0.0% | -10.0% | 75.6 | 43.9 | -31.7 | 0.085 | 0.098 | 0.013 |
| Kelly Johnson | -1.0% | -7.0% | -10.0% | 86.1 | 106.6 | 20.5 | 0.147 | 0.171 | 0.024 |
| Mike Moustakas | 1.0% | -11.0% | -10.0% | 76.1 | 123.7 | 47.6 | 0.149 | 0.186 | 0.037 |
| Khris Davis | -4.0% | -2.0% | -10.0% | 106.2 | 120.6 | 14.3 | 0.214 | 0.258 | 0.044 |
| Leonys Martin | -5.0% | 1.0% | -10.0% | 89.0 | 50.1 | -38.9 | 0.090 | 0.094 | 0.004 |
The two names which stick out in the overall value change are Matt Holliday and Brett Gardner. Both had disappointing 2015 season, but the chances they bounce back in 2016 are not good.
This study is just an initial attempt on the subject and a more in-depth study should be done in the future. In the meantime let know what you think and if there could be any additional information added to the formula.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
Question with regards to McCutchen. Was there any noticeable difference in his numbers for the month of April versus May-October? He seemed to be bothered by a knee injury for about a month and a half which was preventing him from turning on any pitches, so I’m curious to see if April is responsible for him being on this list.
Ohhhhh …. injuries. I can link up my DL database and see if there is any differences for injured players.
Interesting — but what’s the logic for multiplying the decline in contact rate by 2?
The normal range of change in Contact% is about half that of Pull%. Sorry for not adding it.
Really cool stuff, Jeff.
It would be interesting to take this concept and run an ANOVA amongst the 4 groups + control for o-zone contact rate AND z-contact rate separately, being that o-zone contact % is the one peripheral that drops off the cliff dramatically for older players. We’d get an idea about the extent to which overall wRC+ is affected by o-zone contact decline. Hardy’s 2014 and 2015 o-zone contact rate was much lower compared to recent years.
Good idea, let me look into it.
Moose, for all intents and purposes, broke out in a big way last year and stopped pulling because he no longer had to. He showed a huge bump in RC+ and actually increased his contact rate.
In cbssports leagues, Gardner was top 30 overall most of the year.
A very good return considering the low draft pick, even more if you benched him last couple months.
Can you use the reverse analysis as well: Notable increase in contact and pull rates = Players heading toward breakout?
I agree with Paul D. in that a reverse analysis would be interesting. But I’m curious why you used pull% and contact% as precursors to a declining wRC. I’m not doubting they are, but why not choose swst% and gb%? Or any other combination?
hardhit% seems like a good one
I’m seeing a lot of guys who got better on that list. 12/37 improved wRC+ by double digits. 16/37 declined by double digits. (Setting the arbitrary error bars at 10 wRC+ just because.)
It seems to me that injuries need to be factored in. In the cases of Holliday and Gardner, injuries slowed them. If they can return to full health (and maintain it over the course of a full season–which, for Gardner, seems to be problematic), then these declines would not be predictive