It seems like I have gotten a lot of questions lately about how my rankings work, as well as more readers interested, so here’s a basic explanation. I’ll link to it on the Rankings page as well.
Most golf rankings are wrong. They are wrong for several reasons. They assume tournament golf is played against a course. It is not. It is played against a field. You hear this all the time when people will say things like “Justin Rose has been under par in 9 of his last 10 rounds” while “Tiger Woods has only broken par once in the last few weeks.” The problem is, this is completely meaningless. Rose played in three tournaments where the winning score was -18, -15 and -10. Tiger played the US Open, where the winning score was EVEN par. Now, Rose played well, but even playing that well in the US Open probably would have meant 2 or 3 rounds over par.
The second big area that I see that is wrong(and this happens with a lot of rankings systems) is they are created to essentially objectify human biases. In other words, they take what is already assumed to be true by general knowledge and use a computer to fairly rate all teams/people. This is the main problem with the Official World Golf Rankings. While they assume that the majors mean more because they are majors, I ask why?
With those two problems in mind, I had the following goal when creating these golf rankings:
-Start with zero assumptions.
-Create an objective system that rates a player compared to the average PGA Tour field.
The process is simple. Golf scores, as a measure of skill(in the long run), are normally distributed. That allows us to get a standardized score from every round. So, let’s say Tiger Woods shoots 69 in the Masters, when the field averaged 72. The standard deviation for the field is 3. Tiger’s round score would be (69-72)/3=-1. Similarly a player who shot 75 with a field average of 78 and same standard deviation would have the same score. It doesn’t really matter what you shoot, but how the field did when you shot it.
That brings up another issue, is every field the same?
The answer to that is, no. The second step is assigning a field strength to every round played. I take the standardized round score above, and add in the raw field average from the past two years. This varies widely. Something like a WGC(usually strongest fields in the world), PLAYERS or weekends at Major(Aside from PGA and lesser extent Masters, there are enough random qualifiers that get lucky to get into fields that first two rounds of majors are basically an average tour stop.*) will give around a -.25 adjustment to the field, while the first round of the Turning Stone Championship will be around +.1 to whatever the raw score is.
The final step is the tour adjustment. A player with a .00000 ranking would be playing at a level of the PGA Tour average. But, there are three tours accounted for in my rankings. Unlike the world golf rankings, which assume that the other tours are worse(and by how much) my NW and EPGA adjustments are based on comparative rounds over the past 2 years. For PGA to EPGA this generates about 6000 rounds, while its around 4000 rounds for PGA to NW. In individual cases there are varying degrees of success overall, but when you take the whole world of golf it seems these are pretty good adjustments. Obviously, as the standard the PGA Tour recieves no bonus for the Tour adjustment. However, in general players are about .2 worse on the PGA than EPGA, and .35 worse on PGA than NW. So, I add those factors to the adjusted fields above.
So, a players ranking is:
Standardized Score+Raw Field Average+Tour adjustment.
For Tiger Above it would be:
What this says, is against a field of average PGA Tour players, Tiger would be 1.25 standard deviations better than the mean. The same 69 against a field of the Nationwide Tour players at a weak NW event might look like: -1+(.05)+.35=-.6
Using a 69 at the US Open and comparing it with a 68 at the Travelers is like comparing apples to oranges. With the above calculations, all scores are in the form of apples.
If you’ve seen my rankings at all, you will know that there are two components. Ranking and standard deviation. The rankings are the average score of all rounds calculated like above for a two year period. The standard deviation is the standard deviation of all rounds played by a player over a two year span. Interestingly, as a rule of thumb, better players tend to be more consistent. While the average tour player has a ranking of .000 with a standard deviation of 1, Tiger is more like -1 with a standard deviation of .9.
My rankings are not weighted to any recent performance. That’s not saying it is completely meaningless, but it’s hard to say what value it has. From what I’ve found in limited research the relative importance of recent tournaments varies widely and is only fractionally more important than random tournaments. Same applies to course form.
The rankings above represent a players score in relation to the average PGA Tour player and his consistency or standard deviation over one round. To convert it to four, just multiply by four (or three or two, etc.) The standard deviation for multiple number of rounds I found by running some simulations. For the average player(.000,1) it is around 2. I also found that these were proportional to whatever a player’s current standard deviation is. So, for four rounds, multiply the standard deviation by 2. For three, it’s around 1.7. The best way I’ve found to come up with the odds, is set a winning score and adjust it until all the sum of all players chances to win are 1. Obviously, for in-running odds, I standardize the rounds already played and proceed from there.
Planning for one big contenders/pretenders/darkhorse post as well as updated rankings on the entire field tomorrow.