How to work out a price by the spread price ?

In my experience 2.7 mid price is the equivelent to 2.0 unders or overs
and moves about 10% per 0.1 goals, for example 2.80 would be 1.90 overs,
2.90 = 1.81, 3.0 = 1.72, 3.10 = 1.66 etc.
With hindsight 1.78 seems a big price for your example.

It's generally calculated use poisson distribution, once you have the expected goals and supremacy you can calculate the expected goals per team from there you use poisson to calculate the probability of scores below 3 goals i.e. 0-0 1-0,0-1,1-1 add them all up and work out the decimal odds for under 2.5 goals.

poisson distribution should work for total goals.

It doesn't work for correct scores.

P(x; μ) = (e-μ) (μx) / x!  always seems to work for me as long as I have properly calculated the value of μ. Failure to do so can lead to very nasty losses as can wrongly disallowed goals, and incorrect offside decisions.

Amendments following Bortkiewicz should be avoided, I feel, as they only serve to complicate matters unnecessarily.

In-running variations of μ are possible but take a great deal of time. Often, by the time it is recalculated, someone will have scored another goal so buggering up the value of x and the whole issue is screwed up.

Any ideas for a spreadsheet which dealt with the calculation of μ relative to the time of scoring of the first goal would be gratefully received.

Rocketfingers Joined: 20 May 10
Replies: 222 10 Jul 10 22:02 
Uruguay v Germany over 2.5 goals 1.78 on the spreads it was 3 - 3.2 meaning the spreads predicted 3.1 goals. Just wondering how you can work out a thru price from it.


Just taking 3.1 is a bit dangerous. When you look at expected goals per team, the spread can be something ridiculously large like 0.4-0.6; Using only this information, you can see that the correct price is somewhere between those values according to the company, but a small deviation from the middle can make a significant difference to the odds you work out from it.

Investor, not sure what you are meaning. A difference of 0.4-0.6 between the
protagonists would be in line with an 11/10 chance IMO. A small deviation from
middle would not make much change, if at all, unless I'm missing something.
I make lots of comparisons throughout the domestic season and my calculations
usually finish up close to prices offered, although not always. I can't see much
downside the mid price, with perhaps an adjustment to take in the tendency for
punters to buy more often than sell.

Let's say an accurate goal expectation is 0.45. If the spread is 0.4-0.6 and you take 0.5, that's 10% out. As I work out odds across markets by making sure they are in line with eachoter, a mistake like this would be compounded if I was to take it as the basis for further calculations.

For me using spreads to check the accuracy of my models is only slightly better than completely useless.

Investor, I presume when you say 'goal expectation' you mean 'superiority'.
If so, IMO the the difference between 0.45 and 0.5 is the same as the difference
between 11/10 and 23/20 in fractional betting, which is not really that
much, not enough to be concerned about if you give yourself a decent margin
for error.
Regards your view on the usefulness of spread prices for checking the accuracy
of your models, I would remind you that they (the spreads) bet to their figures,
either side, in other words, they put their money where their money is.

Correction ! 'they put their money where their mouth is'.

Using spread prices for comparison is now outdated. Asian prices heavily affect the prices and given certain firms' preoccupation with 'getting short' and closing sellers/winners, the markets are not traded on 'the bid' in any size, leaving them inefficient and inaccurate for the purpose mentioned.

No, I do mean goal expectation (in this case for the (far) weaker of the two teams). Using it in combination with other spreads to price up total goals, over/unders, correct score markets etc. isn't viable.

Using superiority may give you something more useful, but I think the match odds market itself will be a far more accurate indicator of the correct price.

If you DO mean goal expectation then your example of 0.45 (0-4-0.6)
puzzles me. Goal expectation for most domestic football matches is
usually between 2.2 and 3.2, so are you referring to 'in running'
spreads ?