Regression!  For me, it's something of a nemesis.  I've never fully understood it, let alone applied it successfully in the football trading world.  Hopefully, that'll all change today!

The second part of the Bradley Terry model from Andrew Mack's Statistical Sports Models In Excel book centres on regression analysis.  First, what on earth is 'regression'?  Good question!  Regression tells us the correlation between two real-world variables.  In our model, for example, we would like to understand to what extent the 'logistic function' (calculated in the last video) explains margin of victory.  In others words, does the fancy measure we have created actually explain real-world performance.  And, to what extent?

If the answer is 'very much so', then we have an interesting model!  First though, we have to understand a regression analysis report since, without interpreting this, we can't see if the model is any good.  It's not easily interpreted, however, since it's littered with statistical terms.  ANOVA, anybody?  Sum of squares?  In the video, I do my best to explain the terms and point out the most important features of the report.

Finally, we replicate the analysis that Mack demonstrates in the book for our Premier League model.  Does our fancy 'logistic function' measure really explain the results of football games.  Well, kind of ...

Tools And Techniques

- Regression Analysis
- X and Y Variable
- Hypothesis
- Line Fit Plot
- R Square
- Residuals
- Sum Of Squares
- Standard Error