CONCACAF WC Qualification Odds

[voros]

Howdy,

As part of a larger overall project, ran a series of simulations of the entire qualifying process for CONCACAF (and the AFC) based on ratings I derived from an increasingly complex and difficult to explain system.

Anyhow I figured I'd share the results on the eve of the start of qualifying for anyone who is interested.

I ran two groups of simulations, 10,000 runs a piece and recorded which teams qualified by taking the top three spots in the hex, which teams took 4th in the hex and qualified for the Asia playoff, and which teams qualified for the hex. I did this twice because two scenarios are yet to be determined: CONCACAF qualifying in which Guatemala participates, and qualifying where they do not. Rankings:

Qualify for World Cup in Germany 2006 (Including Results of Asia Playoff) :

Team W/GUAT WO/GUAT
Mexico 97.94% 97.88%
U.S.A. 94.88% 94.35%
Honduras 67.28% 75.00%
Costa_Rica 61.14% 68.71%
Guatemala 13.16% 0.00%
Canada 8.02% 10.50%
El_Salvador 6.52% 5.61%
Jamaica 6.14% 5.77%
Trinidad_&_Tobago 5.92% 5.83%
Panama 2.92% 2.61%
Cuba 0.44% 0.69%
Barbados 0.02% 0.03%
St._Kitts-Nevis 0.01% 0.00%
Haiti 0.01% 0.01%
St._Vincent_&_Gren. 0.00% 0.00%
Belize 0.00% 0.00%

Those were the only countries to register non-zero results (the bottom two register below 5/100ths of a percent).

Those numbers for the USA and Mexico seem awfully high to me, but I suppose with neither team ever coming close to going through 10,000 qualifying cycles it's hard to say. Both benefitted from the draw that hurt Costa Rica, Honduras and others.

According to the simulations of the playoff series, CONCACAF would expect to win the Asia playoff 64.40% of the time if Guatemala is allowed in, and 67.00% if Guatemala is thrown out. The only thing the difference means is that the times Guatemala would have finished 4th, are replaced predominantly by teams a bit better than Guatemala.

Chances of Making the Hex (CONCACAF Finals):

Team W/GUAT WO/GUAT
Mexico 99.91% 99.90%
U.S.A. 97.39% 97.44%
Trinidad_&_Tobago 79.15% 79.58%
Honduras 78.22% 88.16%
Costa_Rica 69.50% 78.67%
El_Salvador 39.93% 40.13%
Jamaica 33.78% 33.22%
Guatemala 28.13% 0.00%
Panama 26.72% 26.73%
Canada 21.52% 28.34%
Barbados 8.31% 8.45%
St._Kitts-Nevis 6.70% 6.64%
St._Vincent_&_Gren. 5.66% 5.07%
Cuba 2.57% 4.40%
Haiti 1.93% 2.27%
Nicaragua 0.25% 0.33%
St._Lucia 0.24% 0.17%
Belize 0.03% 0.12%
Surinam 0.03% 0.31%
Dominican_Republic 0.02% 0.03%
Bermuda 0.01% 0.03%

In the 10,000 runs, Mexico failed to make the Hex only 9 times, evidence of a relatively easy run until the hex.

Now is a good time to note Honduras' better chances of qualifying and making the hex than Costa Rica, despite being slotted for the same semi-final group. Why? Because while Honduras has a fairly easy 2nd round qualifier, Costa Rica will almost certainly be facing a not trifling Cuban team. Those runs where Cuba managed to pull the upset were enough to tip the scales in favor of Honduras. Also note that T&T has a better chance of reaching the Hex than either (when Guatemala is included) due to their draw.

It's nice to see a small country like St. Kitts and Nevis post a reasonable outside chance at making the Hex {and as you see above they have a 1 in 10,000 chance of making the World Cup :) }.

Chances of being the playoff team (4th place):

Team W/GUAT WO/GUAT
Honduras 20.05% 25.19%
Costa_Rica 17.02% 20.45%
Guatemala 11.89% 0.00%
Trinidad_&_Tobago 9.08% 8.43%
Canada 8.04% 11.12%
El_Salvador 7.84% 6.91%
U.S.A. 7.71% 9.06%
Jamaica 6.87% 6.27%
Mexico 6.38% 7.46%
Panama 4.27% 3.83%
Cuba 0.66% 1.10%
Barbados 0.10% 0.07%
Haiti 0.04% 0.07%
St._Kitts-Nevis 0.02% 0.03%
St._Vincent_&_Gren. 0.02% 0.01%
Belize 0.01% 0.00%

This team would have missed the WC in 2002, but now has a bailout. The sims say that Mexico, USA, Costa Rica and Honduras would win a sizeable majority of their playoff encounters with the Asian rep, and even Guatemala and Canada would win more than half of the time. In all of these lists, Panama and Cuba look to be the longshot sleepers.

For political buffs here are the chances of the Asia playoff consisting of the USA and one of the following teams:

Chance of USA playoff vs:

Iraq 1.17%
Iran 0.41%
North_Korea 0.14%
Vietnam 1 in 130,000

[Cweedchop]

Very interesting and cool stuff there voros..

[Jimjamesak]

Wow! Very well done and I must say that I'm definately taking more math classes in college.

[gohb]

Great stuff! but can you describe the main features of your formula to us non-math-types?

[ElCesar]

You must have a lot of free time in your hands.

[voros]

Originally posted by gohb
Great stuff! but can you describe the main features of your formula to us non-math-types?
I will, but I need a day or two.

And well, I don't have a bunch of free time, but this is part of my training, or at least that's my argument and that's what I'm sticking to.

Plus, with modern computers, running 10,000 sims isn't too difficult or time-consuming.

[beineke]

Seems like the rating system has a problem with Jamaica ...

[voros]

Actually, they are ranked 7th in the confed which I think is fair (they're about equal with Canada and El Salvador). Where I think they run into problems is that although they got the better draw in terms of making the Hex, in an odd way, they're at a disadvantage to Canada in that if Canada makes the Hex, they will have dispatched two of their main obstacles to qualifying (two of Guatemala, Honduras and Costa Rica).

In other words, Canada only has to play over their heads over a period of 6 games while Jamaica needs to do so for 16. Strangely enough, for the sole goal of qualifying for the World Cup, Canada's draw ain't that bad. If Canada can make the Hex, they have a fairly decent chance at qualifying.

[beineke]

Originally posted by voros
Actually, they are ranked 7th in the confed which I think is fair (they're about equal with Canada and El Salvador).

In the past two World Cup cycles, Jamaica reached the Hex both times, easily. They finished 3rd and 5th.

Canada finished last in the 98 Hex and didn't come close to the 02 Hex. El Salvador finished 5th in the 98 Hex and didn't come close to the 02 Hex. These performances aren't comparable to Jamaica's

So why doesn't the computer system reflect Jamaica's strength? Take a peak at Jamaica's record in games when they have nothing at stake. It's terrible.

[paul51]

INTERSTING INDEED, BUT THERES NO WAY CUBA CAN UPSET THE TICOS, IN THE WORST OF THE SCENARIOS THE COSTARICANS WOULD TIE IN CUBA OR LOSE BY ONE GOAL DIFFERENCE, AND IN SAPRISSA THE Y WILL GET SLAUGHTERED. SIMPLE COMMENT. NICE PROJECT,THOUGH.

[voros]

Originally posted by beineke
So why doesn't the computer system reflect Jamaica's strength? Take a peak at Jamaica's record in games when they have nothing at stake. It's terrible.
The system does adjust for that, though, plus their record in Friendlies and Gold Cup games don't look much different to me than their record in qualifiers

Clearly Mexico, USA, Costa Rica and Honduras are the top 4 teams in CONCACAF and then after that there's a grouping of teams. System says Guatemala is a touch better than Canada who is roughly even with Jamaica who is a touch better than El Salvador and then a gap to Panama, T&T, Cuba and Haiti. The system doesn't like those 5-0 losses Jamaica likes to be on the wrong end of to good teams when they face them away from Jamaica, so maybe that's part of the issue.

Jamaica didn't exactly cruise by El Salvador, they finished with 12 pts to El Salvador's 9, splitting the head to head. The difference is that Jamaica was able to beat Honduras at their "stadium" where they're a bitch and a half to beat, and El Salvador was not able to win its home match against Honduras. Jamaica beat El Salvador 1-0 in Jamaica and lost to them 2-0 in El Salvador.

[voros]

Here are who the system says are the top 64 teams as of 2/16/04:

1 - Brazil
2 - Holland
3 - France
4 - Argentina
5 - Czech Republic
6 - Spain
7 - England
8 - Germany
9 - Italy
10 - Portugal
11 - Ireland
12 - Denmark
13 - Turkey
14 - Croatia
15 - Romania
16 - Belgium
17 - Uruguay
18 - Norway
19 - Mexico
20 - USA
21 - Sweden
22 - Serbia and Montenegro
23 - Russia
24 - Paraguay
25 - Colombia
26 - Greece
27 - Australia
28 - Cameroon
29 - Japan
30 - Israel
31 - Senegal
32 - Poland
33 - Peru
34 - Costa Rica
35 - Ecuador
36 - Chile
37 - Nigeria
38 - South Korea
39 - Ukraine
40 - Honduras
41 - Finland
42 - Slovakia
43 - Switzerland
44 - Morocco
45 - Austria
46 - Bulgaria
47 - Ivory Coast
48 - Tunisia
49 - Iran
50 - Wales
51 - Egypt
52 - Georgia
53 - South Africa
54 - Slovenia
55 - Hungary
56 - Scotland
57 - Iceland
58 - Bosnia-Herzegovina
59 - Bolivia
60 - Belarus
61 - Venezuela
62 - Guinea
63 - Albania
64 - Guatemala

So that's:
CONCACAF = 5
CONMEBOL = 10
UEFA = 34
CAF = 9
AFC = 3
OFC = 1

[beineke]

Originally posted by voros
Jamaica didn't exactly cruise by El Salvador, they finished with 12 pts to El Salvador's 9, splitting the head to head.

That's a bit misleading. With two games to go in this group, Jamaica had 12 pts to El Salvador's 3. After they had clinched their place in the Hex, Jamaica didn't get another point.

Jamaica's head-to-head record in qualifiers, 1998-2002

Mexico 1-4-1
*USA 0-1-3
*Costa Rica 1-3-0
*Honduras 2-2-2
---------------
*El Salvador 2-1-1
Canada 1-0-1
T&T 2-0
St. Vincent 4-0
Barbados 2-0
Surinam 2-0
(* record includes a game that Jamaica lost when they had nothing at stake)

They've been competitive against three of the big four and are undefeated in 15 meaningful games against the rest. They sure look like a top five Concacaf team to me -- they're also #5 in the Elo Ratings, and in FIFA, they're #4.
-----------
More remarks...
-- It's pretty clear that sometimes Jamaica doesn't compete ... in addition to losing their final two games in the 2000 semis, they lost four straight to wrap up the 2001 Hex. They've also lost their last three games in Mexico by a combined 15-0.

-- I disagree with your comment about Jamaica's record in friendlies. Since WC02, I believe they've had five friendlies against teams that reached that event:
0-1 LOSS to Brazil in England
0-0 DRAW in South Africa
1-1 DRAW in Japan
2-0 WIN over Uruguay
3-2 WIN over Nigeria

Pretty good results, if you ask me.

[MoRado]

Here are my odds for the qualifing rounds....

MEX, CR, USA in germany
HON or JAM will play against the asian team
REST OF THE COUNTRIES no chances at all

[JG]

Originally posted by beineke
That's a bit misleading. With two games to go in this group, Jamaica had 12 pts to El Salvador's 3. After they had clinched their place in the Hex, Jamaica didn't get another point.

This is tough to judge...Jamaica did win their first four games to qualify easily, but that reflects an accident of the schedule as much as anything else...it's quite possible that they would have lost in Honduras and El Salvador even if those games had meant something.

Jamaica's head-to-head record in qualifiers, 1998-2002


They've been competitive against three of the big four

Yeah, and it's not like other teams have a much better record at the Azteca.

they lost four straight to wrap up the 2001 Hex.

All of these were by one goal...I don't think they failed to compete in these games. I do agree though that Jamaica seems to be underranked, probably due to their propensity for occasionally losing by a big margin.

They've been doing well in recent friendlies, but there were some ugly losses between the last two world cups.

[TopDogg]

Originally posted by MORADO
Here are my odds for the qualifing rounds....

MEX, CR, USA in germany
HON or JAM will play against the asian team
REST OF THE COUNTRIES no chances at all
I should tell the CSA to save their money on the upcoming qualifying campaign, because you have obviously hit the nail on the head with your scienntific analysis.

[voros]

Originally posted by beineke
-- I disagree with your comment about Jamaica's record in friendlies. Since WC02, I believe they've had five friendlies against teams that reached that event:
0-1 LOSS to Brazil in England
0-0 DRAW in South Africa
1-1 DRAW in Japan
2-0 WIN over Uruguay
3-2 WIN over Nigeria

Pretty good results, if you ask me.
Right, which is why I disagreed with your point about their bad record in games that didn't matter. Certainly friendlies fall into that category and they've played as well in those as they have in qualifiers. It's not like they're ranked that poorly, it just so happens that there's four CONCACAF teams tightly grouped from teams ranked 60-80 and the system likes Guatemala a touch better. Canada and Jamaica are as close to a tie as the rankings can get. Guatemala's like 64th, Canada is like 68th and Jamaica is like 69th.

I could ratchet the rankings so that teams that get blown out aren't punished as much, but I have no information to suggest that such a change would improve the rankings. I think the fact that Jamaica has a bunch of 5-0, 6-0, 9-0 (!!) type results in the last eight years ought to be a negative as clearly sides closer to the top teams rarely lose by those sorts of score lines.

[Razcal]

Seriously, what are the formulas and methods that you have used?

I'd actually like to see the formula and the math you've used to figure this out, humor us and assume that we're intelligent enough to possibly follow along.

I've always thought it a bit dubious when someone comes up with this miraculous system to create a prediction, but is unable to explain how they've done it.

[beineke]

Originally posted by voros
Right, which is why I disagreed with your point about their bad record in games that didn't matter.


But this point did not refer to friendlies in general -- from what I remember, your system doesn't count friendlies at all. It refers to games where Jamaica's opponent had more incentive than Jamaica did:
1) World Cup qualifiers that were played after Jamaica had qualified for the 2002 Hex (lost two of two)
2) 2002 Hex games after Jamaica were out of serious contention (lost three of three).
3) Jamaica's washout in qualifiers for the 2002 Gold Cup


I think the fact that Jamaica has a bunch of 5-0, 6-0, 9-0 (!!) type results in the last eight years ought to be a negative as clearly sides closer to the top teams rarely lose by those sorts of score lines.

Perhaps, but Jamaica's economic circumstances (very poor country, widely dispersed players, minimal domestic league) make it much harder to consistently put a strong foot forward. If they get generally competitive results (I've shown that they do), they should be rated accordingly.

Looking at the 1998 Hex, I think your system (again, based on my possibly faulty memory) goes too far in punishing one bad result:

3. Jamaica 14 pts -5 goal diff.
4. Costa Rica 12 pts +1 goal diff.
5. El Salvador 10 pts -5 goal diff.

Practical consequence
Jamaica finishes ahead and goes to France; Ticos and Salvadorans watch on TV.

Goal Difference (which is all your ratings consider, IIRC)
Costa Rica is an above average team, while
Jamaica and El Salvador suck.

Now let's look at these teams' results in Mexico City:
Mex 6, Jam 0
Mex 5, El Sal 0
*Mex 3, Costa Rica 3*

*Played after Mexico had already qualified

Final standings without that game ...

3. Jamaica 14 pts +1 goal diff.
4. Costa Rica 11 pts +1 goal diff.
5. El Salvador 10 pts 0 goal diff.

All of a sudden, the goal differences line up, and Costa Rica's is consistent with their place in the standings. In this case, I think the standings were right, and the goal difference was dramatically wrong.

[beineke]

Originally posted by JG
It's quite possible that they would have lost in Honduras and El Salvador even if those games had meant something.


Possible, yes, but these are clearly sub-par results for Jamaica, especially the two-goal loss in San Salvador. In 1998 qualifying, Jamaica achieved draws in both El Salvador and Honduras. In those games, both teams had something at stake.


It's not like other teams have a much better record at the Azteca.

Depends how you're counting. If you place a lot of weight on goal difference, then Canada's losses by two and four goals are vastly "better" than Jamaica's losses by four, five, and six.


All of these were by one goal...I don't think they failed to compete in these games.

Granted, "failure to compete" is too strong a term. But if he were concentrating, would Tyrone Marshall have given away the ridiculous PK that handed the US a late game-winner? Also, in their home loss against Costa Rica, Jamaica didn't bother to field anything close to their first team.