Given that his team qualifies again, what are a World Cup squad member's chances of returning? Overall, the answer seems to be about 40%, although this depends quite a bit on a player's age. According to a model I fit, a 30-year-old field player has a 25% chance, a 25-year-old is 50-50, and while there isn't much data on youngsters, they seem likely to stay in the picture. (Note: to produce these estimates, I fit a logistic regression using all field players from the 20 squads that appeared in both the 1998 and 2002 Cups.) Here's what I find intriguing: comparing each country's predicted number of returnees with the number it actually brought back in 2002 ... Code: nation prediction actual france 7.84 13 argentina 8.16 11 usa 6.16 9 belgium 6.18 8 tunisia 8.81 10 paraguay 7.97 9 italy 6.99 8 croatia 7.15 8 england 7.37 8 mexico 7.58 8 saudi 8.82 9 denmark 6.27 6 cameroon 11.25 10 germany 5.30 4 skorea 8.74 7 brazil 7.05 5 japan 8.47 6 nigeria 8.57 6 spain 7.98 5 safrica 8.35 5 The two most notorious failures of the 2002 World Cup brought back more players than expected, while the three semifinalists who returned all brought back fewer. While it isn't clear how strong this pattern is -- the US did well with more returnees than expected -- it will be interesting to see whether it persists in 2006. Also, it's unclear whether a high number of returnees would indicate a coaching error or a shortage of emerging talent ... considering the overall strength of Argentina and France, I'd expect the former, but this is all pretty speculative.
I remember being shocked that Argentina would take Caniggia over Saviola last time. I bet if you fine-tuned this, you could figure out how many more or less players would be expected to return based on the team's preformance.
I think that's a good idea, although it seems like it could be fairly difficult, since a team's performance probably ought to be judged relative to its expectations, and I don't have a sense for how to do this. I'd also like to adjust for the size of a country's player pool, but don't have good ideas for that, either. And finally, I'd like to adjust predictions for each player's playing time, but that information would require some real work to assemble.
This is a cool analysis. But having more returnees than expected can be interpreted in various ways. It could be because our players stayed healthy and effective, a good thing. But it could also be because of lack of competition from new blood. In the case of the three semifinalists that you refer too, it could be that they had a tremendous infusion of new talent.
how did you get your prediction number? was it taking their age and applying them onto a 25% to 50% scale?
By fitting a logistic regression model, I got estimates for the probability of returning for players of every age (which really did come out very close to 0.25 for 30-year-olds and 0.50 for 25-year-olds). These are predictions about whether that individual will come back. In order to get a prediction for the team, all I did was sum together all of these probabilities. Make sense?
I haven't done anything with the 2006 squad lists yet, but here are the model's predictions based on the ages of players on the WC02 squads. Each prediction estimates how many of a country's field players will return in 06. To give an example, because Mexico's players in 2002 were relatively old, the model predicts that only 6.59 of them will return this time. In reality, 6 Mexican field players made it back, so in this case the model looks pretty successful. I'm sure that in other cases, things won't be as tidy. Code: nation prediction mexico 6.59 france 6.67 poland 6.70 argentina 6.74 italy 6.78 portugal 6.90 croatia 6.91 usa 7.03 ecuador 7.28 germany 7.41 paraguay 7.42 sweden 7.74 tunisia 7.85 spain 8.05 costarica 8.12 korea 8.49 brazil 8.51 saudi 8.73 japan 9.09 england 9.21
OK, U.S. had 7 predicted. Actuals are 9 - McBride Reyna Keller Donovan Beasley JOB Pope Lewis Mastro As we knew, Arena is a bit conservative in his player selection ...
Planetworldcup.com has put up its squad lists ... looks like the US actual number is 10 -- Keller doesn't count (no GKs) but Dolo and Wolff do. Code: nation prediction actual sweden 7.74 12 france 6.67 10 usa 7.03 10 paraguay 7.42 10 croatia 6.91 9 ecuador 7.28 9 italy 6.78 8 germany 7.41 8 korea 8.49 9 brazil 8.51 9 japan 9.09 9 costarica 8.12 8 mexico 6.59 6 portugal 6.90 6 england 9.21 8 saudi 8.73 7 poland 6.70 4 argentina 6.74 4 tunisia 7.85 5 spain 8.05 5 ... a little surprising to see France bring back a lot of players again.
Hmmm, my counting was a bit sloppy, although I did leave Dolo off on purpose, didn't realize that he counted according to the rules. As people have commented elsewhere, it's a bit strange that a nation that fashions itself as the fastest-improving soccer country in the world keeps taking more players from the past World Cup than expected, while established powers like Spain, Argentina, and England are taking fewer players despite having such a deep group of existing veterans.
John R, I disagree. A country with a deep talent pool should be taking back fewer (once you're slightly past your peak somebody else will catch you) while the shallow talent pools should bring back more (no competition to the established players).
So far, it looks like a repeat poor showing by the teams who brought back substantially more players than expected. Against the Czechs, the USA started 9 returnees (including Keller) and subbed in 2 more. BTW, with Berhalter replacing the injured Cory Gibbs, the official US returnee number climbs to 11.
Code: nation prediction actual sweden 7.74 12 usa 7.03 11 france 6.67 10 paraguay 7.42 10 croatia 6.91 9 During the group stage, there was another poor showing for the teams with a higher-than-expected number of returnees. In addition to the list above, the Czech Republic apparently returned 19 players from their Euro 2004 team, and after a bright start, they obviously performed much worse than expected. Code: nation prediction actual poland 6.70 4 argentina 6.74 4 tunisia 7.85 5 spain 8.05 5 Results for teams with few returnees were mixed. Spain and Argentina won their respective groups, while Poland and Tunisia were poor. At this point, I tend to believe that having too many returnees is a genuine warning of trouble ahead. In a case like Argentina '02, I think it was a cause of the team's failure; in other cases, it may be more of a symptom.