Patrick Stevens bio photo

Patrick Stevens

Former mathematics student at the University of Cambridge; now a software engineer.

Email Twitter Github Stackoverflow

So yesterday the Wimbledon tennis tournament was decided. The system for verifying whether the tennis ball is out or not (and hence whether play for the point stops or continues) on the main courts is as follows:

  1. The ball lands.
  2. The linesperson keeping charge of the line nearest to the landing point of the ball works out whether the ball landed inside or outside the region demarcated by the line.
  3. The umpire decides whether or not to overrule the linesperson’s decision.
  4. The Hawkeye ball-tracking system determines whether the ball landed inside or outside the region demarcated by the line.
  5. If either player disagrees with the official decision (that is, if the linesperson called “out” when the player thought the ball was in, or the linesperson was silent when the player thought the ball was out, or if the umpire overruled a decision that the player thinks was correct) then that player informs the umpire that ey wishes to “challenge” the linesperson. In this instance, the Hawkeye reading is consulted (and the ball’s trajectory slowly animated on a big screen, for added tension) and regarded as definitive.

The problem I have with this system is the process of “challenging”. Each player starts out with a challenge count of three. If a player makes a challenge, and Hawkeye contradicts the official call, then the challenge count is maintained at its current level. If a player makes a challenge, and Hawkeye agrees with the official call, then the challenge count for that player is decremented. A player cannot challenge if eir challenge count is 0. On entering a tie-break, each player’s challenge count is incremented.

This resulted in an unhappy event in the last match of the Wimbledon tournament. The player who went on to lose (Novak Djokovic) used up his three permitted challenges in unsuccessful attempts to overrule the official rulings. Then in one particularly close game, Djokovic was denied a point when his opponent’s shot was deemed to be “in”. He became angry (displaying the unfortunate tendency of professional sports players to throw temper tantrums at the drop of a hat) and shouted at the umpire that the call should be overruled. He had no challenges remaining, and so could not force the official decision to be reassessed; I suspect his attitude very much did not help to press his case at this point. Later, the commentators showed the Hawkeye ruling to the TV broadcast; the opponent’s shot was in fact “out”, and Djokovic was vindicated. As I say, he went on to lose (pretty comprehensively, I gather, although I didn’t really pay attention); it is conceivable, though admittedly unlikely, that this dispute cost Djokovic the match.

My question is this: why do we rely on linespeople to do that which is done better by Hawkeye?

Would it not be massively more sensible if the linespeople were allowed to do exactly what they normally do (as a salve to those who do not wish to sully the tradition), but the umpire were provided with Hawkeye’s ruling after every point so that ey could overrule as necessary? This changes nothing except the umpire’s ability to carry out a task ey already has to do. Of course, in instances where Hawkeye is unavailable, such as on the lower courts at Wimbledon, nothing need change.

Hawkeye supposedly has an average error of 3.6mm, roughly equivalent to the fluff on the ball. I propose that the umpire should be provided with the possible error along with Hawkeye’s decision, and that it should be down to eir judgement which verdict to accept in such tight cases that Hawkeye might have made an error. (I would suggest defaulting to the normal method of judgement in that case - that is, “continue as if Hawkeye had not been invented”.)

The only reason that I can think of to limit the number of allowable challenges is to prevent time being wasted in the administrative process. However, umpire overrules (and challenges themselves) happen rarely enough that I think the following procedure would be quite sufficient:

  1. The ball lands.
  2. The linesperson keeping charge of the line nearest to the landing point of the ball works out whether the ball landed “in” or “out”.
  3. Hawkeye determines whether the ball landed “in” or “out”.
  4. The umpire reads Hawkeye’s decision off a screen.
  5. The umpire decides whether or not to overrule the official call.
  6. If the umpire decides to overrule the call, the ball’s trajectory is animated slowly on a big screen.

Now, this does (of course) do nothing to resolve the problem of conflicting verdicts during a very fast rally - the umpire cannot concentrate on both the game and the Hawkeye reading at the same time. But then there’s no existing solution to that problem anyway, and I do not propose to resolve this problem at the current time.