I know ya’ll been wonderin’ which soccer team is best. Traditionally, them Europeans rank their teams by points accured during the season: 3 points for a win, 1 for a draw and zilch for a loss.

Now a group of German physicists says this ain’t the best way to do it. After analysing all the results from a German football league called the Bundesliga since 1965, a total of 12546 games, they reckon that goal difference is a better indicator of which teams are best.

Andreas Heuer and Oliver Rubner from the University of Munster in Germany say that a team gets the same number of points whether it wins 1-0 or 6-0. But surely the bigger win is a better result and should be rewarded accordingly. And the statistics bear this out.

Of course, what these guys haven’t appreciated is that leagues are more about entertainment than team ranking, whether the sport is soccer, basketball or tiddly winks. The points system is designed to handicap the best teams, so that the competition remains open for as long as possible. Otherwise, the league rankings would be more or less determined before the first half of the season were over.

More interestingly, Heuer and Rubner look at a number of footballing rules-of-thumb to see whether they hold true under statistical inspection.

For example, they establish that every team has an advantage when playing at home but there are no teams where this advantage is stronger than others. Instead, teams who more often win at home are simply better teams.

And they also have an interesting analysis of winning and losing streaks. The statistics show that a losing streak makes a team more likely to lose again. But counterintuitively, a winning streak of more than 4 games also makes a team more likely to lose in the near future.

The researchers also found a puzzle. The statistics show that teams perform equally well or badly throughout the season but there is a step change in performance over the summer break. This can account for about 40 per cent of a team’s overall perfomance.

The other 60 per cent varies over a much longer timescale of 20 years or more. So good teams remain good over a period of 20 years or more.

This timescale is much longer than the period individual players and coaches are involved with a club. So what factors are responsible for maintaining peformance over this period?

They’re scratchin’ their foreheads in Munster over that one as I write.

Ref: arxiv.org/abs/0803.0614: Fitness, Chance, and Myths: an Objective View on Soccer Results

This entry was posted on Friday, March 7th, 2008 at 6:04 am and is filed under At the seaside. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.