Thursday, 14 October 2010

Recalculating


The first order of business is to correct my mistakes from yesterday. First, I wrote that the percentage of draws in the EPL last season was 33.8%, but the true figure was 25.8% (the 33.8% was the HT draw). Over the past ten years, the draws have averaged 25.8%, so last season was pretty average, although 2009-10 did start off with a distinct lack of draws.

But there was also a more serious error in my P/L calculations, as pointed out by Simon M, which meant that the profits, and ROI, were nowhere near as high as I had reported for most categories.

The revised numbers show that backing the draw in every game at the average odds would have resulted in a loss of 47.5 points over the season, an ROI of -12.5%, but there was one class of match which returned a profit of 9.11 points (ROI 38%) and that was where the home team was in the Strugglers group, and the away team was a Top 4 team. Just 24 matches, so a small sample and more research needed here.

Backing the draw in intra-group matches at average odds showed a small loss of just. 1.87 points, ROI of 2.23%, but if you hunted around for the best odds, this becomes profitable at 0.24 points and 0.28%.

Matches where the away team was two grades higher resulted in a small loss of 0.15 points from 48 matches, an ROI of -0.32%.

What does all this mean? For a start, it means that finding the value by backing the draw isn't as easy as it appeared to be yesterday, but when I think about it, those numbers really were a little unlikely!

However, given that I am working with average prices and assuming a 5% commission, both of which can be easily bettered, by being selective, we are still coming out ahead.

Prior to a month ago, I don't have a record of what the strong draws predicted in the EPL were, but it's almost certain that none were from the higher grade v lower grade category.

The four draws in the EPL that I do have a record of are all in the -1 category, i.e. the home team was one grade worse than the away team, for example this weekend's Fulham v Tottenham Hotspur, or between teams of the same grade. The Elo ratings don't recognise grades though, just ratings, so if Liverpool's and Everton's troubles continue, it is quite probable that Liverpool v Blackburn Rovers will become a candidate.

Coming up this weekend, there are five intra-group matches, but the ratings have only one of these as a draw, and one as a strong home win. We do have a -3 contest which would have been profitable last season, (Blackpool v Manchester City) and the ratings have this as a weak draw.

Fulham are currently unbeaten incidentally, a fact that hasn't received too much publicity, with 6 draws from 7 matches so far.

The ratings should prove to be better than the grades as a selection method, because they are more fluid. Assigning a grade at the start of the season is problematic, not for the top teams, but for the bottom half of the table where fortunes can change rapidly from one season to the next. It's not too long ago that Middlesbrough and Portsmouth, for example, were solid teams.

For the record, this weekend's intra-group matches are:

Bolton Wanderers v Stoke City
Newcastle United v Wigan Athletic
Wolverhampton Wanderers v West Ham United
Everton v Liverpool
Blackburn Rovers v Sunderland

2 comments:

Anonymous said...

Hi.
I hope your research on laying odds on in the bundasliga was correct, as its a simple system i have decided to add to my portfolio.
This season is going ok so far though.
Thanks.

sam said...

Hi Cassini. I have recently started trading and have set up my blog. I would be grateful if you could add me to your long list of blogs too :)