‘LUCKY’ ROBINS DEFY THE ODDS, POTTERS PAY THE PRICE...
THE table never lies. Isn’t that the old cliche? There’s just one problem. Data analysis suggests the Championship standings tell more porkies than our current Prime Minister.
Take, for instance, Hull City. Grant McCann’s side were tipped for relegation in August, so their current placing of 12th indicates a job well done. Yet if the Tigers had achieved the results their performances warranted, they would actually be sitting in third place.
That, at least, is according to Expected Goals, a method of analysis that predicts a team’s long-term performance based on the quality of the chances they create (xG) and concede (xGA). Expected goals assigns a probability of a chance being scored ranging from 0 (impossible to score) to 1 (certain to score). This is based on an analysis of historic data which examines the relationship between the type and location of an effort at goal and the eventual outcome. Put simply, a shot from eight yards out in the centre of the goal has a high probability of finding the net. A header from the same position does not. Similarly, a longrange strike from 30 yards would have a probability in the region of just 0.02 (two per cent), whilst a penalty kick clocks in at 0.78 (78 per cent). Ultimately, the aim of xG is to prove the old adage that form is temporary and class is permanent by exposing unnatural hot or cold streaks. Like any statistical analysis, xG is not perfect. It cannot account for intangibles like confidence and team spirit, nor the heroics of a world-class goalkeeper. As the performance of Hull indicates, there are evidently factors that confound the system. But as a rough guide to the ‘real’ performance of a football team, it is an interesting tool. So what does it say about this season’s Championship?