Friday, 22 April 2011

Pick Your Poisson

I was hoping the gentle dig in my last post might generate an indignant comment or two, and Blue Orange stepped up to the plate with:

To get my odds for the unders/overs markets of Premier League matches I look up the form of the last 5 home/away games of both teams, look at what they've scored and conceded and using the poisson distribution come up with each teams chances of scoring goals against their opponents. This takes a long time and I certainly am not a guesser. If I judge something to be value going by my ratings then I'll have a bet. Out of the 10 matches per week only on average 2 show any value.
I had suspected Poisson distribution may be a factor actually, based on the fact that often Blue Orange's calculated price and mine are in the same ballpark and my ratings spreadsheet has evolved to calculate prices on all major markets including the under / overs. Once you have the probabilities of Team A scoring and Team B scoring, it's easy enough to calculate the odds on any score or market.

The tricky part is, of course, accurately calculating those probabilities in the first place. I am confused with the statement that "out of the 10 matches per week only on average 2 show any value" since the only way there can be no value is if your calculations of the probabilities exactly match those of 'the market'. If not, then if value doesn't exist on one side, then it exists on the other - i.e. it's a value back or a value lay.

The screenshot above shows my prices for the Liverpool v Birmingham City match this weekend, although not with last week's results included as I am still not yet caught up from my vacation. This is still a work in progress, and the numbers are quite often wildly out of line with the market, but it illustrates the point that there are many opportunities to use the results of your calculations to find value in matches.

Mark left a comment too on my line "I mean seriously, how can someone in London possibly have any edge on a match in, say, Argentina, Bulgaria or Croatia? It's hard enough to find value on your local team - never mind a team from an obscure league playing on a weekday afternoon."
A little harsh perhaps when one of the most sucessful, proven non-league (English) betting experts is Skeeve. He is Croatian and still lives in Croatia as far as I know.
Firstly, Croatia was picked simply because the country begins with a 'C' and was not intended as a dig at any one individual. I have seen the name Skeeve on one of the blogs I link to, but have no idea who he is, where he lives, or what he specialises in. It's certainly an unfortunate name! There's nothing wrong with specialising (far from it), or betting on games in Argentina, Bulgaria or Cyprus (better?), but the point I meant to make is that many people happily dip into leagues they know nothing about simply because there's little action that day. How often do you see a tip for a game in Bulgaria on a Saturday?

I just did a little digging and found some more details about Skeeve, and evidence that specialisation pays perhaps:

Who the hell are you, Skeeve?

I'm a non-league betting advisor. I'm 30 years young, studied sociology, worked as a journalist, a copywriter and a screenwriter, but then became Skeeve full-time and haven't regret it so far. I could also tell you who are my favorite writers, directors and musicians, but that would be kinda off-topic, wouldn't it?

Are you any good?

Take a look at the stats - my long-term ROI is 14%. Both good and bad runs are perfectly normal, but I always have a last laugh. Hard work never gets unrewarded, especially in the long run. And other irritating catch-phrases.

What are your stakes? Are you using a 1-10 units scale?

I'm currently using a 1-5 units scale and my average stake is a bit more than 3.5 units (advised bank: 80 points).

When are you usually sending the picks? Do I have to be online 24 hours a day?

6:30 PM Friday and 11 AM Tuesday (or 7 PM Monday if there are no midweek picks, which is the case most of the time).

So, you're covering the Blue Square leagues, right?

Right. The service focuses on non-league markets that are strongest in terms of availability and liquidity - Blue Square Premier asian handicaps and all kinds of Blue Square Premier and Blue Square South doubles.

Why Blue Square leagues? Why not Premiership, Championship, Swedish hockey, Ukranian volleyball?

Because I've been following the English non-league world for quite a few years now and the fact is that bookmakers can't keep up with all the non-league news (transfers, injuries, suspensions, work commitments, financial problems... It's never boring in the non-league world) all the time - it's either that or some terrible odds compiling. I'm a hard worker and a long-term winner and that's something I wouldn't want to change, so I'll stick with the Blue Square leagues for now - those I know best.

Are you making these stats up? Who can guarantee these are actually the right stats?

I was posting my non-league picks & previews for free between December 2006 and October 2007 – on my old blog, on the Betting Advice forum, on the ex-Punters' Paradise forum (I was a resident tipster there), on the Laka Lovica forum (I was a moderator there)... All the picks since the service became paid (October 2007) have been independently monitored and verified, by the highly rated Betting Advice surveillance team from October 2007 to May 2010 and by the superb Secret Betting Club since March 2008.

What about your picks? Do they come with a match preview or are they just plain picks?

All of my picks come with a detailed match preview. I'll even find out where you can find the best odds, second best odds and in fact any odds that in my opinion have value at that moment.
And there's more. I like his answers - and his English is a lot better than my Croatian. if you're interested.


Mark James said...

I hadn't actually realised you'd used Croatia as one of your obscure countries/leagues. I was just pointing out it is possible to specialise in an obscure foreign league.

Unfortunately I think Skeeve is a closed shop at the moment. Limited membership to protect prices.

Anonymous said...

With regard to the comment about only 2 from 10 games on average are value. I only bet when the Betfair odds differ from mine by 12%. Anything lower than that ends up having a pretty low margin once Betfair have taken their commission. I hope that helps.