Thought I’d weigh in on this interesting question posed on Paul Kedrosky’s Infectious Greed blog.
While Tiger Woods may very end up winning the tournament, how do these performance surprises happen? What is it about golf, in other words, that makes competency so fragile?
To put some numbers on it, Tiger is deemed the best golfer in the world because he wins 28% of all major tournaments that he enters. Meanwhile a top tennis player, like Roger Federer, wins a significantly higher percentage of majors, with Federer at 35% and the great Bjorn Borg at 41%. Winning competency in tennis is more stable, it seems, than in golf.
What is golf so fragile? The ball doesn’t move, and the player isn’t in motion, so two major variables disappear. Granted, the golf swing is longer and the clubs change, which is not the case in tennis (there are no serving rackets, etc.), but staying in place with the same equipment and a stationary ball are all major advantage.>/p>
The answer seems obvious to me, and several of the comments get it right. In a typical round of golf, Tiger Woods has somewhere between 62 and 80 opportunities to exercise his skill against the overall distribution of opponents who each have between 62 and 85 opportunities to exercise their skill as well. In the course of a tournament, this is around 280-300 opportunities. In contrast, a typical tennis match gives Federer or Nadal anywhere from 400 to 700 opportunities to exercise their skill advantage. Over the course of a tournament, this may be upwards of 3500-5000 opportunities. If you consider Woods’ or Federer’s advantage as something akin to the house advantage at a casino, a small edge becomes more definitively pronounced the more bets placed; the confidence of a larger sample size in revealing level of skill is higher than that of a smaller sample size.
Furthermore, in golf, variance in a particular stroke creates dependent events; one bad shot can slightly increase the likelihood of additional bad shots for the scope of that hole (one of 72). In tennis, the scope of dependent events is limited to a single point (one of perhaps 250-300). While smaller than the other effect, this also contributes.
The other factor: within a tournament, tennis matches are played serially and one-on-one, typically progressing from seeded matches where the house advantage is larger to ones where it is smaller. Golf is every player going against a relatively small sample against a long tail of competitors—something similar to the birthday paradox comes into play as the likelihood increases of finding someone capable of stringing together 4 outstanding rounds (relative to their expected advantage) as the number of simultaneous competitions goes up.
Finally, Kedrosky implies that prior performance above the norm in close proximity to the tournament in question should predict future performance. On the one hand, it should be easy to dismiss this. If separate tournaments are independent events, then hitting 14 of 14 fairways the week prior should have no bearing on whether he does so this weekend, no more than the roulette wheel coming up red 14 times in a row should affect the 15th spin. However, you can’t discount this entirely. Good performance is an indicator of lack of impediment (e.g., lack of injury) so to the extent that the course and conditions are similar from week to week and there being no evidence of injury occurring in the intervening week, we at least know that some of the factors that could degrade from his optimal advantage are not present. Put another way, rather than Tiger being “in the zone”, we may only really infer that he’s “not not in the zone”.
Further related reading: Intensity of tennis match play
