Insight: A lower win probability when hitting down the middle does not necessarily mean you should hit down the middle less often.
In Part 1, I show Murray and Raonic had higher win probabilities at net partially because they tended to approach the net on easier incoming balls. In other words, they were more likely to win the point because they were already at an advantage. Given that, the opposite should also be true. If a tactic is usually used when at a disadvantage, then that tactic should have a lower win probability. However, just as a higher win probability for an offensive tactic can deceive, so can a lower win probability for a defensive tactic.
A prominent defensive tactic is hitting down the middle, which reduces errors and takes away the opponent’s angles. Per Table 1, in the 2016 Wimbledon Men’s Final, Murray hit 48 of his 185 baseline-to-baseline rally shots down the middle, and Raonic hit 53 of his 206 such shots there. To clarify, missed shots count toward these totals, and shots missed anywhere but wide may be classified as down the middle. In addition, “t-1” refers to the incoming ball, and SHORTt-1, CENTEREDt-1, and DEFENSIVE t-1 are control variables. As in Part 1, I use regressions to determine the importance of the controls.
First, I regress WIN on CENTERED, which is the new decision variable, without controls. As shown in Table 2, the win probabilities of Murray and Raonic were 14.0% and 10.7% lower, respectively, when they hit down the middle than to a corner. Also, when I combine the samples, the difference in win probability, now 12.3%, becomes significant at the 5% level. Therefore, the trend for each individual player is most likely real but statistically underpowered. However, when I add the controls, all three of the differences in win probability shrink, and none are significant at any level.
Ultimately, these results can be interpreted in multiple ways. On one hand, even after those differences shrink, they remain negative and large enough to matter. On the other, that does not mean Murray or Raonic should try extra hard to avoid hitting down the middle. For one, I may have omitted an important control, such as how far back they stood on each shot. In time, I may look deeper into this question. But for now, the most consequential takeaway is this: when a tactic is notably defensive or offensive, its win probability may understate or overstate its effectiveness respectively.
Table 1: Summary Statistics
Table 2: Win Probability Regressions
