« Competition in public services | Main | Thatcher's "axe" »

June 04, 2010


The Silent Sceptic

Surely a man of your talents would realise that this analysis is total hoop. I've just done a quick simulation of these games between 6 teams, assuming a generous average of 4 goals per game. Each team's probability of scoring was uniformly distributed across the 90 minutes of each of their 521 games. Pure random chance showed that the gritty 'team 5' scored 18 times in the last minute, showing their never say die attitued, while 'team 6' didn't have the same tendency to go down fighting - scoring only 6 times in the last minute.

It's well established that Monte Carlo methods are poor for estimating low probabilities, and so it is in this case. The stuff about penalties stands up to scrutiny much better, as the probability is further way from the extrema. However, the conclusion about penalities isn't a national stereotype, it's just an observation.


Completely spurios stats. Think about the sample size and then adjust further for the natural churn in players and managers.


I'm not entirely sure it makes sense to think of efficiency as about scoring in the last minute. Surely "efficiency" is about getting the job done early.

However, I have to disagree with KaKTy3. The natural churn of players and managers has been taken out of the equation by looking over a long period (200+ games), necessary to see only the effects of "cultural differences" between footballing nations.

Tim Almond

"And one other thing: Germany really are better at penalties. In shoot-outs since 1960, they have scored with 94% of their penalties, whereas England have scored with just 50% and Italy with 65%."

In 1990, 7 of the 16 teams were involved in at least 1 penalty shoot-out. In 1994, 6 of the 16 teams after the group stage were involved in a penalty shoot-out. But Glenn Hoddle didn't even bother practising in 1998, despite such high probabilities. Maybe he thought that applying time to on-field play would mean they could outright win each game rather than facing the possibility of penalties?

photo ex machina

Some would say it's no surprise that stereotypes are often true - after all, that's why they are stereotypes in the first place!

Fred Kapoor

After reading the article I was kind of speechless and I tried to look at this in the complicated way, however, just like @photo ex machina puts it: " stereotypes are often true - after all, that's why they are stereotypes in the first place!" and I couldn't agree more.


silent sceptic - you are neglecting the importance of the prior hypothesis.

If you have a prior hypothesis that team C is a never-say-die team then it is meaningful if your small-sample survey supports that hypotheis.

Open -ended research 'I wonder if any of these teams....' is different.

The comments to this entry are closed.

blogs I like

Blog powered by Typepad