1-1 Mexico – South Africa
0-0 France – Uruguay
Before getting into the sims, it’s worth pointing out the basics here. Because of the two draws, 5 points is a magic number now. A win and a draw in the next two games for any of the four teams in Group A puts them through to the knockout round. Considering the closeness in quality of the top three teams in the group and the host country being the other team, advancement of any kind by any method is probably something any would sign up for in a heartbeat.
Under Domenech, France has been a fairly low scoring team and this tends to cause a lot more draws than normal. This makes them a team that tends to be stronger in knockout rounds than in group stages since the winning team gets more than twice the value of a draw in the group stages. Theoretically Mexico have hurt themselves by only taking one point from the hosts, but hosting really is greatly beneficial for whatever the reason.
There’s very little change in the sim results, and most of it is within the range of random chance. However the 6% increase in the chances South Africa has to advance and the balance of the 6% decrease for Mexico probably is meaningful. Mexico’s expected points from this one was 1.78 while South Africa’s was 0.94. France’s was 1.46 and Uruguay’s was 1.24. So while South Africa only got a slightly greater number of points than expected, all three of their opponents got less with Mexico leading the bunch.
So obviously there will be bigger movement in the odds to come, but it’s nice to have the tournament finally here.
6 responses so far ↓
1 ac // Jun 12, 2010 at 3:24 am
Am I right in saying your match %s don’t incorporate SA’s home advantage? And therefore SIMs not quite accurate?
2 Voros // Jun 12, 2010 at 4:31 am
No South Africa’s home field is accounted for, though it’s worth considering how accurate the adjustment is.
But yes they would be much longer underdogs without that adjustment.
3 ac // Jun 12, 2010 at 4:42 am
Thanks. Far be it for me to question your workings but I can confidently suggest your home advantage is considerably too low and would have a significant impact on your numbers.
Great site by the way, only just came across it.
4 Rasmus // Jun 12, 2010 at 12:37 pm
Very nice. Soccer seems to be very hard to predict accurately (I have some background in prediction modeling, though never for sports), looking at how both your and Nate’s model are never confident enough to assign a team more than a 80, maybe 85% winning chance, even in the most lopsided matchups.
I would really love Italy and New Zealand to play each other 20 times after the world cup and look at what the results would be. My bet would be that Italy wins at least 18-19 games.
5 Yair N // Jun 12, 2010 at 8:48 pm
Very interesting site, thanks.
I was wondering if after you post each result and how it changes things, you could post what is the probability the model gave to the actual result (e.g what probability did the model give to Korea winning 2-0 or to Argentina winning 1-0).
Is that (or something similar) too difficult to calculate?
6 Voros // Jun 13, 2010 at 9:36 am
Yair,
No that’s easy and you can do it yourself if you like.
Use the following formula:
Where
n = 33
k = number of goals scored
p = predicted goals scored/33
If you don’t know what the first term is, it’s a symbol for a combination. If you’re using excel just use their function COMBIN(). Otherwise the formula for a combination is
n!/k! * (n-k)!
With the ‘!’ being the universal symbol for a factorial.
Leave a Comment