Thus far, we have focused on the technical side of our World Cup simulator - namely, how it has been created.
Today, we are going to look at individual odds for the tournament. We can use our methods for creating the simulator to see which individual markets appear to be good value.
Our earlier article outlined how we found equations to predict the odds of a match between any two World Cup teams. These were:
- To win: 1.7135R^(0.4623), where R is the team's coefficient divided by their opponent's.
- To draw: 1.7958R^(0.2114), where R is the outsider's coefficient divided by the favourite's.
We used the group stage matches to create these equations, which in turn led to calculating probabilities. We now work backwards - we use our equations to calculate the probability of each outcome for each group match. We can then use the following equation to give an 'expected return' if you bet £1 on that outcome:
- (return if win x probability of win) + (return if lose x probability of lose)
Now, the return if you lose is -£1 (you lose the pound you bet), whilst the probability of your bet losing is just 1 minus the probability of you winning. So the equation can be changed to give:
- (RP) - (1-P)
For example, for the opening match of the tournament we estimate Russia's chances as 69.8%, and if they win you'd have earn 31p (not including your stake). So your expected return from putting £1 on Russia to win is:
- (0.31 x 0.698) - (0.302) = -0.085 = -£0.08
In other words you'd be expected to lose 8p for every £1 you put on the match. So it doesn't seem like a good idea to bet on Russia to win.
Doing the same calculation for every outcome of every group match, we can identify which matches are good value. Below you can cycle between the groups to see our identified best bets:
Groups A & B
Groups C & D
Groups E & F
Groups G & H
Similarly, there's second value when Iran play Spain and Portugal. In these cases a draw looks good value as well.
It is fairly unusual for the favourite to be overpriced, but it appears that Spain are marginally overpriced against Morocco. This may be because the bookmakers expect Spain to have qualified by then.
It's a similar story for Group D, with one obvious exception. Iceland to upset Argentina appears to be very good value. What a story that would be!
Elsewhere, a draw between England and Belgium and a Polish win over Senegal are possible avenues to explore.
In our previous article we gave probabilities for where each team would reach in the tournament (e.g. there is an 8.77% chance of Belgium finishing second).
Therefore, we can repeat the same analysis we have done with individual games above to show the expected return of betting on all possible team/outcome combinations:
This is because both sides are in a group with Tunisia and Panama, two weak sides. So they have a very high chance of getting into the knockout stages, and with a weak Group H they also should also both make the quarter-finals. From there, it's just one game to reach the semi-finals. Of course, it's still unlikely either side will reach the semis (33% for Belgium and 29% for England), but these probabilities are higher than the odds would have suggested.
Elsewhere, Denmark and Nigeria are good value for a second round (or even quarter-final) appearance, thanks to the openness of their groups.
However, what you may also notice is how overwhelmingly negative the table is. This is because odds are stacked in bookies' favour, so on average this should be negative.
To summarise, these are the bets identified as best value:
- Saudi Arabia to beat Russia
- Egypt to beat Uruguay
- Iceland to beat Argentina
- Panama to draw or beat Belgium
- Panama to draw or beat England
- Belgium to reach the final or win the tournament
- England to reach the semi-final, final or win the tournament
- Denmark to make the last 16
- Nigeria to make the last 16
- Sweden to make the last 16
All of these are unlikely to happen, taken individually. However, there is a good chance some of them will happen, and their odds are such that they will reward you disproportionately well.
I should also add that I take no responsibility for any lost bets! Our similarly-identified matches in 2014 would have brought a profit, but similar methods on the last two General Elections backfired. It's almost as if it's a gamble.