Friday 19 October 2012

Me vs. Global Betting Conglomerate

A few posts back I posted my attack and defence rankings for Gameweek 8 based on my model and later posted the equivalent odds from a well known bookmakers that shall remain nameless. In this post I am going to briefly compare and comment on the key differences between my attack rankings and those of William Hill's [sic]. 

GW8 Attacking Rankings - me vs. Bookie

The line is the best fit through the data points.  Firstly, I am pleased of course that there is a fairly decent association.  Nothing is majorly wrong with my model when compared to the million of dollar backed bookies model.   Woot me.

So let's look a the differences.     Anyone above the line is favoured by the Bookies to score more than I have them down to.  Similarly, a team below the line I favour to score me than the bookies do.  There is a cluster of teams around Chelsea, QPR, Southampton etc. that  are all much of a muchness.   At the top and bottom end there is agreement on Liverpool and Reading, Stoke.

The key differences, those teams furthest from the line, are of interest.  William Hill have Man Utd considerably more likely to score than I do.  This is due to the market I believe.  Punters will be drawn to a fixture such as MAN vs STO regardless of the fact that Utd are not in top gear and Stoke are defending well.  

Don''t get me wrong, I do expect United to score in this game, but not to the same extent as the bookies do.  Essentially the bookies have Stoke as 7/1 to keep a clean sheet,  I have them more like 3/1, meaning in 7 repeated games the bookies would give Stoke 1 clean sheet in 7 and me 1 in 3.  I expect the bookies are protecting themselves against the large number of bets that will be placed on Utd rather than against. 

In the same vein, I have Norwich as more likely to score than the bookies do.  This one puzzles me a little.   Arsenal aren't known for a great defence away from home, although this year things may be improving in this respect.  Bookie have an Arsenal Clean Sheet at 8/5 whereas I have them a lower (more unlikely)  value of ~11/5, therefore I would bet on Norwich to score if I could find that bet.

Disclaimer:  The only bet I ever made was Everton to beat Nottingham Forest 2-0 in 1997. And they did.  But I lost the ticket (only a £1). I am not a well-versed gambler but I do think this analysis has some value to fantasy mangers.

1 comment:

  1. Although I only acted on "intuition" I had Norwich likely to score, naming Holt in my punts for Gameweek 8.

    The problem with trying to analyze the bookies' numbers, as you note with the United game, is that they often fix the numbers based on where they believe the most number of bets will be placed.

    Interesting analysis, nonetheless. Keep up the good work.