Occasionally, someone like Rory McIlroy will win a big tournament, then miss the cut in another big cut the following week. You’ll hear the big media raise there hands up in the air proclaiming they have no idea what is going on anymore in the world of golf(like they ever did). The real question is, does how someone played last week really matter in the first place?
To take a preliminary look at this I looked back at 2009. I compared a players performance from last week, with his overall ranking and one random tournament, to see how much of an influence last week’s tournament had on this week.
The answer, in short, not that much.
As you can see, the r-squared and correlation values are slightly higher, but given the standard deviations from tournament to tournament there’s a decent chance that any correlation could have been random.
Obviously, there are some problems with this. Namely that I only looked at the 30-some-odd tournaments from 2009, and I only used last week and not last tournament played in to make it easier for me. It also doesn’t address the issue that the last month of play might be more meaningful than just one tournament with varying participation.
Still, I think it’s pretty safe to say that last week’s tournament is not all together a great predictor of success in the present week.* As a point of reference, I picked out Tim Clark as the winner of the PLAYERS.
*Although, to be fair, 4 rounds of golf is such a small sample with high variance that 1 year rankings were not a great predictor, either. My one year rankings have had a r-sq of between .8-.85 with something like the moneylist, so they do well overall, but one tournament is an incredibly random event
Overall, Tim Clark is about .44 standard deviations(1.28 strokes) better than the average PGA Tour player. That would make him among the best players in the field this week with an approximately 18-1 chance to win if you weigh in all rounds over the past two years equally.
While, I struggled with a way to weight last week’s tournament, I eventually settled on multiplying last weeks results time 1.58(last weeks correlation and r-squared, divided by random tournament. I have no idea if this is a good way to do it.)
When you do that Clark becomes about .46 standard deviations(1.32 strokes) better than the field, all other things equal. That improves his chances to a whopping 16-1 to win against this week’s field.
This was a far from deep study of meaningful past performance, but it’s safe to say that any meaning last week had is almost certainly overrated by the general public.