Has anyone tried modelling the outcomes of the Euro 2016 or any other sports event? i.e. assign risk/return attributes to each team as if they were stocks (with all the fundamentals and all…) and compare with the bookmaker’s coefficients to find discrepancies.
Or do you assume the bookies are fully efficient as well as rigged which collapses the market (re: overround)?
I had done one in 2010 South Africa World Cup but that was long before I was into CFA and modelling in general. There I used only bookies statistics, which was a looping pointless model. I could try doing another one midway through the games with a narrower field and “more tools”…
Given that some bookies allow “shorting” teams as well as “marking to market” I say it is mimicking the financial market ever more.
Since bookies’ coefficients will eventually converge to betting exchanges coefficients, you are essentially betting against other people, some of whom may use quantitative tools, so it will be whether your model has more predictive power.
On another note, I see two problems with Euro:
-
Lack of historic data. Even if you manage to quantify isolated performance of players using their club season(s), there is no guarantee performance for clubs will translate into performance for national teams.
-
Small breadth. There are too few events to bet on, so you won’t even know whether your model works (doesn’t work) or you’ve just been lucky (unlucky).
If you’re serious about it, I think it’s better to focus on leagues - many games, more historic data. Cup competitions are c**pshoot.