Tuesday, 10 May 2011

Double Chance?

Gold All Over has it's first (non-spam) comment which means at least one person is taking a look at my screen-shots over there! The Football Analyst recently concluded that the DC (Double Chance) wasn't for him, and I need to decide how to play the longer priced selections that the Poisson Value system highlights. Keith's comment was:

Wow that really is a piece of work - looks great!!

How do you decide whether to lay or back the value selection? e.g Stoke v Arsenal - Stoke had a home edge of 91% but in your selection you had layed Arsenal whereas Everton v Man City - everton had 58.7 home edge and in selection you had backed Everton for win - is it a judgement call? Or probably I may be missing something!!!!

Enjoy following you on here and over at "Green All Over"
The answer is that I quite arbitrarily picked odds of 4.0 as the deciding line between backing the selection outright of getting the draw on my side. So for example on Saturday, Everton were backed at 3.0 v Manchester City whereas for Stoke City at 5.5 on Sunday, I went with the lay of Arsenal. After a larger sample than 11, I shall revisit this, but, as I said in my reply, I don't like my losing runs to be too long!

It's a promising start for this system at a tricky time of the season, and it's a little unfortunate that I didn't get all the parameters in place earlier, but at least it is ready to go for 2011-12.


Peter "redimp"Hunt said...


Really love reading your thoughts, any chance you could follow my blog


kind regards


Graeme Dand said...

Hi Cassini.

I've been reading your stuff across at Gold All Over although with such a small sample size, it's hard to draw too many meaningful conclusions at the moment. Expect comments/observations in future. ;)

With regards to the DC point, the only reason I decided I shouldn't track this any longer is that it felt like I was trying to cover my ass a bit too much. The whole point of my ratings is to find value backing teams to win and all my backfitting and backtesting only looked at this.

I then noticed a lot of draws appeared and that meant Draw No Bet became a feasible option for me and I thought that covering the draw was a way to smooth the returns. I then picked up the fact that I seemed to be hitting a lot of draws this season in particularly and historically, this appears to be the case too and therefore, dutching the draw and the selection to win was also profitable. In 2011, backing the draw instead of the selection to win is the most profitable strategy by far, so that could be another strategy to look at!

Someone made a fair comment privately to me that if I wanted my ratings to be judged according to their profitability, then having so many ways to track their profitability wasn't going to help me in the long-run. I think that is a fair call as clearly, Double Chance is nowhere near as profitable as backing teams outright and therefore, why should I dutch on the draw when backing the draw outright is a loss making exercise long-term with my ratings?

Anyway, to cut a long story short, I agreed with them that having 13 systems, with 3 different ways of playing the bets was probably too much for any of the results to be taken seriously (a monkey could find a profitable set of systems from 39 attempts!) and therefore, reducing the number of systems and ways to play the bets was the way to move forward.

I've a lot of work to do this summer to move this on another step and improve the profitability of the ratings and come up with the best way/ways to play the bets next season but I'm looking forward to the challenge!