World Cup: Holland to win, Spain to slump
We know the score. Yes
World Cup Stats Spain is chewing on the chorizo of defeat thanks to Switzerland's 1-0 victory yesterday, and history shows this result is a major setback.
In the last 40 years only one top 10 team has won a World Cup group after an opening game defeat. Only four have finished in the top two. The tournament favourites Spain now face a serious group H struggle to move to the next round.
But should Spain really have been the tournament favourites? They were according to bookies and pundits, but the UK's top soccer statistician has Holland as the top team, after its form in the final set of friendly matches saw it sneak past Spain and Brazil.
Dr Ian McHale is Senior Lecturer in Statistics in the University of Salford. He and his team are also behind the Actim Index, which is the official ratings system of the English Premier League. This is an objective rating of players using official match statistics, not one based on subjective notions such as flair - not to be confused with the Sky-linked Opta Index.
McHale did some modelling of the teams in the lead-up to the World Cup. He used an 'ordered probit' model (an ordered regression model, stats fans) to estimate the probability of the three outcomes of a match - whether the result will be a win, a draw or a loss.
You can mark his working out here in an article he wrote for the Institution of Engineering and Technology. This doesn't include his final number crunch on the World Cup which put Holland top.
The Reg caught up with Dr McHale at half time during the Spain – Switzerland game. At the time he was saying “It's beautiful to watch, Spain do know what they're doing”, and hypothesising that someone like Owen Hargreaves would have been better for marking Spain's Xavi out of the game, than Frank Lampard.
His model only uses the results of previous International games, the number of goals scored, and and location of game. Weighting comes from Fifa's database of team rankings.
Things that favour a team are: if they're at home; or haven't travelled too far; are ranked higher than the opposition; if they're an improving team according to ranking; the seriousness of the game (friendly or major tournament - better teams take major tournaments more seriously); and if they've been winning recently.
“I've got a database of 9,000 international match results over an eight-year period,” he says, viewing his research as a way of showing statistics can be interesting, and seeing whether the “market is efficient”.
Sponsored: Global DDoS threat landscape report