Question about betting lines and 'implied probability' (american odds)

Hi all,

This is my first post here. Thanks for having me as a member!!

In short, I'm trying to understand what exactly 'implied probability' is and how to convert that value to a basic 'probability of the team winning regardless of the sportsbook making money' value.

Say I take some American odds betting line, and convert it to an 'implied probability'. So far so good.

Is there any way I can look at the sportsbooks' probability they've set, and instead convert it directly to some probability that the team will actually win?

Another way of asking this -- how can I remove the juice from the 'implied probability' calculation? Is this even possible?

I'm trying to build a model and I'd like to see how it performs vs the betting lines, but without the built-in profit of the sportsbook.

I hope this makes sense. Thanks everyone!!

 

Hello Mr_Mush:

The following method might help you with your model. Using the lines posted by the book, you can remove the juice and figure basic probability of each team winning the game.

Say have a game Boston +120
Houston -140

implied probability for Boston 100/(100+120)=.4545
implied " " Houston 140/(140+100)=.5833
total of both is 1.0378 (meaning book's juice is .0378)

calculate for Boston- .4545/1.0378=.4379 for Houston- .5833/1.0378=.5620

Boston has .4379 probability to win game, while Houston is .5620 to win. Therefore TRUE ODDS would be Houston approximately -128 (.5620/.4379=1.28)

Hope this helps

 

I prefer to divide the juice equally to every outcome.

-> Boston +120 / Houston -140
--> juice 0.037878

Boston 0.4545 - (0.037878 / 2) = 0.4356
Houston 0.5833 - (0.037878 / 2) = 0.5644

The difference of this method to the proportional distribution of the juice correlates with the odds. For odds > +400 it's quite significant.