Bombadil wrote:Neu Leonstein wrote:Assuming that I didn't mess up the code somewhere, there are a couple of things going on. The first is the ELO scores. I didn't use FIFA ones, both because I think they're bad and because the draw was based on them, so using them to analyse the draw would be a bit recursive. France has a 1984, which is pretty high, but a bit below the top-ranks like Brazil (2131) and Germany (2092). Plus, there are a few teams with similar scores to France in that follow-up pack.
The second issue is that France doesn't have an easy draw. Assuming that they top the group, their next opponent is the runner-up of group D, which is probably Croatia (1853). Then it'll be a choice of Spain, Portugal or Uruguay. And after that Brazil and Germany.
They can win all those matches, but the problem is that to win them all in a row often enough to show up as a significant proportion of total tournament victories they need a number of positive net shocks to happen in a row. It just so happened that when I ran my simulations, they didn't get those very often. Again, assuming that everything worked as intended. I worked through a couple of loops step by step to try and make sure the logic works as it's supposed to, but there are no guarantees. I'm still learning.
Anyway, for what it's worth, I agree with you. France is probably a better chance than the ELO scores give it credit for, and I think Brazil is a substantially smaller chance than the ELO scores imply. But I kept my own views out of this as much as possible.
Ah.. ok.. I don't think there's anything wrong with your math here - but I do think there's an issue with the ranking. As far as I can tell you've used ELO and then to some degree mapped to WC ranking?
The issue is that the South Americans seem very high on ELO and I suspect it's possibly the nature of their qualifying campaign. They seem to have the greatest variance in ELO rankings (Colombia jump 17, Peru drop 21) though for the life of me I can't see how Brazil dropped five and Argentina only one.. I guess friendlies.. Brazil drew with England recently..but it's y-o-y rankings.. still when you adjusted Argentina shoots up in the rankings on your system, which kind of belies their poor qualifying campaign and poor team sans Messi.
I wonder if you can download (not 'you' per se) first team FIFA player rankings on Playstation and rank accordingly.. and then enter the same random shock.
I say 'I wonder' because I don't think the effort of running it is particularly worth it and, to your point, the interest in this is post-tournament and seeing who out-performed or not. Like you say I'm not sure you can easily quantify randomness in football.
Hang on did you just say Argentina have a poor team bar messi?? Are you mad just think about Aguero, Higuain, Di Mária etc?