No football trading model is complete without a prediction mechanism. These do something essential: they generate 'theoretical' or 'implied' odds according to the model; we can then compare these odds to what is on offer in the real world (ie. Betfair) to identify value opportunities.  It's a common process in football trading, and the best in the business are really good at it!

In the video, I build a prediction mechanism for the Bradley-Terry model based closely on the example from Andrew's book, though we do make some minor improvements :-)  Along the way, we apply formulae such as VLOOKUP, the normal distribution and complete the process with a data table to instantly create odds for multiple 'margins of victory'.

So, is this it?  Are we ready for success in sports trading now we have completed the Bradley-Terry model for Premier League data?  Well, I wouldn't say so.  As I explain at the end of the video, the real skill lies in exploiting a model (not simply 'building' it), and constructing a repeatable process around it that allows you to exploit a small edge over a long period of time.

Want to see more from the New Horizons series?  Let me know in the comments under the video, or in the community forum.

Tools And Techniques

- VLOOKUP
- COUNTIF
- Column Chart
- Normal Distribution
- NORMDIST
- Data Table