Saturday, 4 April 2015

Rant And Response

FTL sponsor Ian Erskine is a hugely successful punter, and is a shining example of what can be achieved with a positive mindset and disciplined approach. To the best of my knowledge, he is not a qualified statistician, but he knows his way around a spreadsheet, understands value, and likes to go on the occasional rant.

His latest ranting is in full here, but this paragraph caught my attention:
Those services that back the draw need to overturn a deficit of -341 points, I don't care what they say they are not going to do it with any success. Backing home and away teams in every game again encounters losses both actually almost the same again betting the home team or away team on Betfair in every game. Now that is how hard football is and that is the starting point. To profit you need to select games that can overturn that negative start and then of course find enough of them to make it worthwhile.
The proof-reader may have skipped over the second sentence there, but what Ian is saying is that in the long-term, backing or laying all home, draw or away selections isn't going to be profitable. This should not come as a surprise to anyone of course, and with the draws generally priced with an implied probability of 33% or less, draw backers are on the wrong side of the favourite-longshot bias.

With not too many services backing the draw, Ian may well have had the XX Draws in mind when he wrote that, but looking at the Big Five leagues (EPL, Serie A, La Liga, Ligue 1 and Bundesliga) for the two full seasons from 2012-14 using Pinnacle Sports prices, backing every draw would have resulted in a loss of 253.66 points from 3,652 matches (five matches did not have Pinnacle prices, so the average prices were used).

The Favourite-Longshot bias holds true with 254.44 points lost in matches where the implied probability is 28% or less. A simple filter thus puts the blind draw backer into profit already.

Pinnacle's prices have only been with us (at least on Football-Data) since 2012, so two seasons is a small sample, but as I have mentioned before, matches priced at around 3.15 or below are borderline value at best, and XX Draw qualifiers in this range have been 'unofficial' this season. If we filter out matches where the implied (Pinnacle) draw probability is less that 29%, we are left with 1,279 matches and a profit to Pinnacle prices of 32.30 points (an ROI of 2.52%).

Using Joseph Buchdahl's idea of then looking at the returns when taking the best prices, and the ROI% climbs above 4%. There are those who sneer at such 'meagre' returns, but as Ian says, winning at football betting is not easy, and I suspect not many could beat that return over 1,000 bets.

Once the current season is over, I'll take a look at the numbers for 2014-15. I might also find the time to look at the Lower Four leagues in England. I know Peter Nordsted's Drawmaster is 'band' based between 3.3 (30.3%) and 3.56 (28.1%) in the English Leagues.

This post is is no way meant to suggest that one should blindly back every draw priced in the ranges specified, but it does show that there might be a sweet spot between the under-priced outsiders (longshots) and the under-priced shorties where additional filters can improve an edge. There can be value at any price.

No comments: