Forming an Opinion of Price in Professional Golf Tournaments

To develop an opinion on winning probability in any betting market, you need to be able to rate the chances of each contestant relative to the rest of the field. This is by no means an easy task for most punters, if it were, there would be fewer successful bookmakers. In this article, I describe how you can use publicly available information to develop your own tissue prices for professional golf tournaments.

Sporting bodies the world over attempt to rank participants. From the contender rankings in boxing through tennis to darts, many sports, both team and individual, rate professionals relative to one another. Unfortunately, many ranking methods are somewhat flawed and of negligible use to bettors. Take FIFA’s world football rankings below, as of 3 July 2017.


Spain, Italy and England would likely be short priced favourites against any of the other three teams shown and yet all are ranked behind Switzerland. FIFA’s rankings are flawed because the 211 rated international teams simply don’t compete often enough nor against one another.


The World Golf Rankings

The official world golf rankings (OWGR), however, are mature and well thought out. Here’s why I believe they are robust enough to form the basis of an educated golf model.

  1. They are comprehensive, covering 19 global tours including the feeder tours
  2. Ranking points are ‘decayed’ over a two-year period meaning that most recent performances count for more
  3. Each event includes a Strength of Field (SOF) rating based on who is competing in it. Stronger fields rightly attract more ranking points

On the downside, the golf rankings include some features that don’t necessarily suit the purpose of betting. Most notable of these are the maximum of 52 ranking events in two years and minimum point levels per event, even if it attracts a poor field. For this reason, the rankings can be used as the basis for development of a tissue, but bettors are advised to customise based on individual preferences.

Shown below is a simple way of converting world golf rankings into probabilities/odds. Firstly, here’s the top of the world golf rankings as of 25 June 2017.


The column that we will use to develop our relative rankings is the Total Points column, indicated with the red arrow. Total Points includes all the events that a player has competed in over the last two years, and is not limited by the 52-event maximum that OWGR applies [1].

Like many ratings and regression methodologies, we can take each player’s points and divide that by the number of points of all event participants to calculate an approximate probability. If only the above ten players were competing in a ten-player event, this basic calculation would look as follows.

RankPlayerTotal Points% of totalImplied Odds
1Dustin Johnson567.3417%6.02
2Hideki Matsuyama415.2812%8.22
3Jordan Spieth387.1911%8.82
4Rory McIlroy302.669%11.06
5Jason Day291.139%11.73
6Sergio Garcia309.29%11.04
7Henrik Stenson316.659%10.78
8Alex Noren249.767%13.67
9Rickie Fowler283.578%12.04
10Brooks Koepka285.828%11.95
Field Total3,414.60

The total amount of points in this ten-man field is 3,414, as shown at the bottom of the image. Each of the player’s totals are divided by that number, for instance for Dustin Johnson 567.34/3414.60 is 0.166 or 17%. The implied odds are simply 1 divided by this number, i.e. 1/0.166 = 6.02.

This methodology is fine, however the raw ranking points from OWGR do not accentuate the class of the elite players enough. For instance, at the time of writing there is only a 100-point difference between world number 25 Phil Mickelson and world number 100, Lucas Glover. For all of Lucas Glover’s achievements, he is some way behind Phil in ability and results.

To make the model better reflect this skew of ability, we need to transform [2] the Total Points column. In the example below, the Total Points column has been transformed by raising it to the power of 1.5 (^1.5).


   1.50  
RankPlayerTotal PointsTransformed% of totalImplied Odds
1Dustin Johnson567.3413513.4121%4.78
2Hideki Matsuyama415.288462.7513%7.63
3Jordan Spieth387.197618.7912%8.48
4Rory McIlroy302.665422.768%11.91
5Jason Day291.134967.418%13.01
6Sergio Garcia309.25437.008%11.88
7Henrik Stenson316.655634.689%11.47
8Alex Noren249.763947.166%16.37
9Rickie Fowler283.574775.197%13.53
10Brooks Koepka285.824832.137%13.37
Field Total3,414.6064611.29

Dustin Johnson is now calculated as a $4.78 chance (21%) with everyone below Jordan Spieth drifting. Whilst the example uses ^1.5, the optimum transformation to apply is entirely up to the modeller. One way to determine whether the transformation is improving the model, is to compare the outputs with historical Betfair prices (accessible via data.betfair.com).

At this point, we have created a 100% tissue making no adjustments for known factors [3] and are yet to take the exchange market into account. Whilst our tissue is a good approximate guide, it is readily improved by adding both factors. In the example below, I have merged our model prices with a mythical Betfair win market.

The blue cells are the mythical Betfair odds for this 10-man event, and the orange cells are a 50/50 merge between our model and the Betfair prices. Note the merge is taken from the implied probability (the percentages) not the odds.


RankPlayer% of totalImplied OddsBetfair OddsBetfair %50/50 Merge50/50 Odds
1Dustin Johnson21%4.783.7527%24%4.2
2Hideki Matsuyama13%7.63138%10%9.62
3Jordan Spieth12%8.487.513%13%7.96
4Rory McIlroy8%11.917.513%11%9.21
5Jason Day8%13.01119%9%11.92
6Sergio Garcia8%11.88128%8%11.94
7Henrik Stenson9%11.47157%8%13
8Alex Noren6%16.37234%5%19.13
9Rickie Fowler7%13.53138%8%13.26
10Brooks Koepka7%13.37264%6%17.66
Field Total101%101%

Now that we have a relatively well-formed opinion on probability, we can consider how we might bet the tournament. In the (simple) example above, one might choose to back Matsuyama and McIlroy (both are larger prices than our 50/10 model), whilst laying Dustin Johnson. Correctly staked, we could keep the other seven players as small winners and cheer the field on, almost guaranteeing we have something to watch late on Sunday.


Wrapping Up

To wrap-up, I will demonstrate how you can easily add a margin to the 101% markets developed above to create full-field lay offers (i.e. the bookmaker strategy discussed in the previous article).  The starting point is your target over-round for the market. The higher you make it, the less players you will lay, the lower you make it the lower the likely return on volume in the long-run.

For this 10-man example, the target over-round is set to 110%. Whilst there are more sophisticated techniques of converting a 100% book to one with a margin (discussed in article 4), in this example, we simply apply the margin equally across all players.

RankPlayer50/50 Merge50/50 Odds110 Percent110 Odds
1Dustin Johnson24%4.226%3.84
2Hideki Matsuyama10%9.6211%8.79
3Jordan Spieth13%7.9614%7.27
4Rory McIlroy11%9.2112%8.41
5Jason Day9%11.929%10.89
6Sergio Garcia8%11.949%10.91
7Henrik Stenson8%138%11.88
8Alex Noren5%19.136%17.47
9Rickie Fowler8%13.268%12.11
10Brooks Koepka6%17.666%16.14
Field Total101%110%

To get to our target percentage, we divide the starting percentage of each player (marked a. on the image), by the original market percentage (b.) and then multiply by our target percentage (c.).  For Dustin Johnson this is 0.24/1.01*1.10. Once applied to all players you have created a book from which you can lay the field on Betfair. Welcome to bookmaking golf!


Related Articles

Introduction to Golf Betting

Those of us lucky enough to play golf will understand the wild swings in form our mediocre talent affords ...

Read More

Intelligent Golf Betting: The Derivative Golf Markets

Read More

Intelligent Golf Betting: The Majors

There are four majors each year in golf, along with one flagship event on each of the two main ...

Read More