Wednesday, April 05, 2006

Golf Ratings

JayDee asked a question below about how to go about finding 'value' when betting on golf. I'll have a stab at describing how I am currently going about meeting that challenge. To be clear, this is very much a new work in progress. I have had enough success historically with golf to have set up a dedicated fund. However, this is the first year, and it's currently losing, but the approach may be of interest.

My starting-point is to have compiled "handicap ratings" for every player on both tours based on their results last season. They received a rating for each outing, with a rolling average covering their most recent 20 outings. I then keep this updated after each tournament.

The basic rating is achieved as follows. I come up with an average score for all golfers who make the cut based on the mid-point of all those who made the cut. These golfers scored "level par" for the tournament. If the winner shot ten shots lower, then they get a handicap score of -10, and so on. Those who made the cut but shot higher than the mid-point get a similarly calculated +score. Those who missed the cut get a blanket score of +10.

The ratings are specific to each tour, so those who play on both tours have two separate ratings. This in itself is interesting. Currently, for example, Ernie Els has a US handicap score of -5, and a European score of -10. This is actually quite consistent with other golfers, allowing me to speculate that the US tour is on average roughly 1 shot per round more competitive than the European tour.

The final ratings aren't in essence that much different from the weighted "scoring average" that each tour provides. However, there are advantages in keeping your own ratings, not least the very act of doing them provides you with unique insights that just looking at stats cold never will. Also, the ratings are kept on a spreadsheet that allows me to look graphically at how each player's form is ebbing and flowing, plus at their winning and placing strike rates and clusters.

These ratings form the basis of my pre tournament form study. They are core 'ability ratings'. I then add positives and negatives for very recent form (last 3 tournaments), and for course form, or likelihood of the course suiting.

For pricing up markets, this information allows me to cluster players into groups, to which I allocate probable winning percentages, and then compare to the early bookies odds (as soon as possible after the bookies put their prices up - usually Monday), and hope to pick off any prices I see as out of line.

The ratings in theory should work very well for "head to heads", though in reality it's unusual for the bookmakers to put together two golfers much out of line with each other. They are also very very useful for pricing up the tournaments in progress, particularly at the end of each day's play.

I'm playing with a number of tweaks to the system, as it is a work in progress, but I am happy that this provides me with some sort of edge worth pursuing.


I hope that was of some interest.

1 comment:

  1. Hey jaydee - happy birthday. It's my 44th birthday today also!

    Ice hockey doesn't register over here at all, but I did put a few pounds on Chicago as suggested - and my betfair account suggests they won - nice one, thanks!


    I'll look out for your head to heads, (whilst hoping Pampling lets you down!!)

    Cheers

    ReplyDelete