In this chapter I'll tell you about Miller's average, but I decided to give the chapter a more global name, because I was able at one time download a few articles on maths in betting from a finnish site. Unfortunately all articles are in Finnish. Not actually :) A lot of things are dwelled upon in these articles, from building a prediction model to defining a managerial change impact on team perfomance. Most of the formulae there are rather compicated, but I'm sure some people wil find them useful. Before explaining Miller's average, I'd like to ask a simple, at first sight, question: in some championship team A plays against team B. They have the following average stats: team A scores 3 goals in a game, team B concedes 3 goals a game. Average team scores 1 goal a game. Question: how many goals will team A score? The answer is very much likely to be "3 goals", and the formula "(3+3)/2=3" will be given as explanation. I'm afraid to disappoint you, but A will score 5 goals. I used Miller's average formula: S1 = A1+D2-Av, where S1 - number of goals A1 - average goals for of team A D2 - average goals against of team B Av - average goals for of the whole championship I'll try to explain this formula now. First of all, we should notice that team A has a very strong attack, they score 3 times more than an average team (3 goals to 1), second thing is that team B has very poor defence, they concede 3 times more goals than an average team (3 goals to 1). Now why should A's excellent attack score only 3 goals (that they score against average teams) against B's poor defence? It's obvious, that there should be more goals. This formula is quite simple compared with those articles, but you may be develop it, for example you may consider home advantage or some other factors. |