World Cup 2014: "True" odds bad news for favourites

Belgium, along with Costa Rica, have to be big lays in their next matches

In his fourth instalment in a series which aims to show the value of ratings to the armchair punter, Jack Houghton, a self-confessed know-nothing when it comes to the beautiful game, gives us the "true" odds for all the Round of 16 matches, explains why Brazil might still represent value, and suggests that extra time is a lottery.

With the knock-out stages about to start, the World Cup betting picture becomes much clearer.

The group stages can sometimes be frustrating for the ratings follower - especially the final round of games - as teams may adjust tactics and teams to reflect the position they find themselves in, rather than playing their "normal" game, upon which their rating is based. From here on in, all teams are trying to win (for the most part, at least), and this simplifies things no end.

With this in mind, I thought I'd save you all a bit of time and publish the "true" odds for the Round of 16 games, based on the system of converting Elo ratings to odds that I outlined last time:

Brazil 1.76
Draw 4.00
Chile 5.49

Columbia 2.84
Draw 3.34
Uruguay 2.87

France 2.00
Draw 3.64
Nigeria 4.44

Germany 1.27
Draw 7.69
Algeria 12.05

Netherlands 2.04
Draw 3.57
Mexico 4.33

Costa Rica 3.83
Draw 2.90
Greece 2.53

Argentina 1.93
Draw 3.85
Switzerland 4.52

Belgium 2.92
Draw 3.29
United States 2.82

Comparing these "true" odds to the Betfair market throws up some interesting differences. If we trust in the Elo ratings, then all of Brazil, Argentina, Columbia and France are underpriced: yes, they are the most likely winners, but not by as much as the market suggests. Of most interest, though, are Costa Rica and Belgium. Costa Rica should be 3.83n/a shots, but are 2.546/4. Belgium should be 2.9215/8 shots, but are 1.9420/21. Even working in a significant margin of error to our calculations, both teams have to be major lays.

With regards to Argentina, Columbia and France, I would not put anyone off laying them also; however, when it comes to Brazil, it might be worth leaving that match alone. Over the years I've read countless studies which have quantified the home advantage enjoyed by football teams (endless pleasurable hours are yours should you want to check out what's available on Google Scholar), and those who create ratings will know that home teams automatically have a measurable advantage, beautifully demonstrated in this lovely bit of analysis of European league competitions.

When it comes to the home advantage of World Cups, it's impossible to be as precise, mainly because the sample size is so small, but looking back at my own ratings over recent competitions, it looks as if the advantage may be around 200 Elo points per game (there is a huge range: 200 is the median). Factoring this into Brazil's next match, the "true" odds are closer to:

Brazil 1.30
Draw 6.67
Chile 12.20

Which might suggest that Brazil, currently 1.608/13, are value to win that match in normal time.

One last thing to keep in mind as the World Cup moves to the knock-out stage... When it comes to extra time, Elo ratings are not an especially good predictor of who will win a match. Again, the sample size is small, but my analysis suggests that even a team with a significant rating advantage of 200 Elo points is only around 6% more likely to win a match that has gone to extra time, so watch out for market over-reactions in-play.