Anyone who builds a football model reaches the same conclusion. The odds market is very hard to beat. Here is why, and what we do about it.
The price collects information
An odd opens based on the bookmaker's knowledge. Then it moves: news on injuries and line-ups arrives, professional players bet, prices get corrected. By the close before kick-off the price has absorbed all of that.
Closing line value (CLV)
CLVi compares the odds you took with the fair closing odds of a sharpi bookmaker (after removing the margin). If you consistently get a better price than the close, you are seeing information before the market does.
CLV does not depend on whether the bet won. That is why it needs less data than profit to tell you something. It counts as evidence of skill, short of proof. We track CLV on our tips and publish results on the performance page.
The favourite-longshot bias
Many studies find that odds on outsiders tend to be lower than their probabilities justify, while favourites are relatively cheap. The effect is called the favourite-longshot bias. An analysis by Hegarty and Whelan at CEPR finds clear evidence in 1X2 markets and none in Asian handicapi. Another paper (Daunhawer and co-authors) finds that betting against the bias is marginally profitable, while betting with it loses heavily on average.
Most of the studies we found are working papers and theses, not journal articles. We read them as an indication.
What we do
The policy is simple. When a sharpi market exists (Pinnacle, Betfair), we only play if the odds beat its fair price. An edge that rests on the model alone is not played.
Sources: Hegarty and Whelan, VoxEU/CEPR, Oikonomidis and Johnson, Southampton.