Something Doesn't Add Up

Eggmund returns with a question regarding my numbers from League One for the last five seasons:

Hello, not sure I've quite understood the data.

How can blindly backing home and laying home both lose? Or how could backing home, away, and draw all lose quite so much, when you would expect to be only slightly down if you backed all the options? Is it that the bookmaker's margin is between 1-2%?

How much is the bookmaker's margin on football matches, and does the data imply that the odds are broadly right (less the bookies spread)?

This may be ignorance on my part as most of my activity is on Betfair, where markets add up to 100.2%ish and Betfair takes its (healthy!) cut at the end.

The average over-round for sports-books varies a lot, and from league to league, but the arrival of Pinnacle has helped.

As Eggmund surmises, the over-round is the reason why, in the long run, most 'blind' strategies will slowly bleed money. Over five seasons and 2,760 bets, that number will add up. 

Note that I adjust prices for the over-round when calculating the lay returns to give a realistic picture.

What is not a realistic is the claim from the Secret Betting Club that backing all 2,416 home teams "in the top 4 English Leagues and the SPL last time out" would have returned a profit of 81.7 points at 3.4% ROI**.

Fake news!

I have to point out that it's no secret that the total number of matches in the top four English Leagues plus the top Scottish League last season was actually 2,264, (380 + 552 + 552 + 552 + 228) and using the more realistic and achievable Pinnacle prices, backing all Homes had an ROI of only 1.3%.

Homes were still in the green by 30.16 points, but not close to the 3.4% claimed as an ROI.

I'm not sure I understand the logic of lumping different levels of football together in this way. It may be a 'secret' strategy but each league has its own personality, and League Two or the Scottish Premier League have little in common with the English Premier League. 

As Home backing in those two leagues (Two and Scottish Premier) would have lost you close to 60 points, I'm not sure why they were included at all. It's important to understand where the profits and losses are coming from.

Surely a much catchier lede would have been that backing all Homes in England's top three leagues last season would have been up by:

Stewboss left a comment on my National League post saying that:

The NL picture looks very different if you factor 2013/14 results into the mix Cassini!

How far back you want to go looking for trends is up to the individual. Personally, I think two or three seasons is good, but in the case of the National League, the last quarter of the 2013-14 season did also show a profit on Away teams, a signal easily missed when looking at the season as a whole. 

Free Tip: Trends don't always start with a new season... 

** The Secret Betting Club's numbers were "based on the results as taken from the football-data.co.uk website, which tracks the best early odds from a panel of 3 bookmakers".