ealing kid wrote:
I used to really value the Fink tank stats until the other week when i saw their predictive stats for the World Cup (click on the image to enlarge):
http://www.timesonline.co.uk/article/0, ... 05,00.htmlAccording to their stats:
- France are the favourites to win the world Cup
- Brazil should be 8-1 to win, not the 3-1 offered by the bookies.
- Holland and Germany have (almost) as good a chance of winning as Brazil and germany are actually more likley to reach the final.
- Argentina only have a 5% chance of winning, the same as Sweden
- Italy, Argentina and England are all expected to get knocked out before the quarter finals.
The worst thing is they actually had the nerve to publish this in the paper without any expanation for the results. I think stats are great, but this does show that there are instances where they are seriously flawed.
France favourites? That's a mistake.
Brazil 8-1 to win: perfect. This is a knockout competition where anything can happen, not a round robin. 3-1 is probably too short.
Germany are the hosts and therefore likely to do quite well. Holland surely do have almost the same chance of winning as Brazil.
The one thing that seemed wrong with the stats was Sweden's high chance of winning the thing - in fact they have more chance of winning the group than England (but only just). But if you think about the latter fact it makes sense; Sweden are more consistent against teams like Paraguay and Trinidad and Tobago than England are, and they probably do have a very good chance provided they don't lose to England (what are the chances of that?)
Italy, Argentina and England out before the quarters: by expected you mean "greater than 50% chance" I assume. I don't see much wrong with this if you look at the teams they have to play (apart from England maybe).