Predicting The English Premier League-using stats

[Matthew]

System Stats:

ALL HW D AW No of Bets: 60 10 28 22           Staked: £2050.48 £335.99 £951.74 £762.75 Returned: £1769.31 £321.35 £1005.11 £442.85 P+L: £-281.17 £-14.64 £53.37 £-319.9 Strike Rate: 38.3% 70% 35.7% 27.3% Yield: -13.7% -4.4% 5.6% -41.9%           Avg Back Odds: 3.22 1.49 3.75 3.34 Avg Win Odds: 2.77 1.38 4.01 2.32           Each Person: £37.78

Back in time for the 20th.

[Matthew]

Date Home Away Prediction Odds Bookie Stake Correct Return 20/Oct/2007 Arsenal Bolton HW 1.29 Betfair £34.75     20/Oct/2007 Aston Villa Man Utd D 3.70 Betfair £34.75     20/Oct/2007 Blackburn Reading D 3.60 vcbet £34.75     20/Oct/2007 Fulham Derby D 3.75 Betfair £34.75     20/Oct/2007 Man City Birmingham D 4.00 Betfair £34.75     20/Oct/2007 Middlesbrough Chelsea D 3.90 Betfair £34.75     20/Oct/2007 Wigan Portsmouth D 3.25 Bet365 £34.75

[Matthew]

I backed a couple of these a few days ago. So odds may have changed - I can
prove the odds I achieved if needs be.

The reason for the delay is that SkyCard now charge you a cash handling fee
for placing bets - they didn't used to and the 2% cashback was quite nice!!!

Anyone know of another credit card (with perks) that allows you to bet without
issuing a cash handling fee?

[Matthew]

Date Home Away Prediction Odds Bookie Stake Correct Return 21/Oct/2007 West Ham Sunderland D 3.55 Betfair £31.82

[Matthew]

Originally Posted by Matthew

Originally Posted by Matthew

Complete nonsense...........I hope for better this weekend.

System Stats:

ALL HW D AW No of Bets: 68 13 37 23           Staked: £2325.55 £370.74 £1192.06 £762.75 Returned: £1939.16 £365.67 £1130.64 £442.85 P+L: £-386.39 £-5.07 £-61.42 £-319.9 Strike Rate: 36.8% 61.5% 29.7% 26.1% Yield: -16.6% -1.4% -5.2% -41.9%           Avg Back Odds: 3.21 1.46 3.72 3.36 Avg Win Odds: 2.75 1.37 3.98 2.32           Each Person: £33.2

[Matthew]

Date Home Away Prediction Odds Bookie Stake Correct Return 27/Oct/2007 Birmingham Wigan D 3.50 Betfair £30.54     27/Oct/2007 Chelsea Man City HW 1.57 Betdirect £30.54     27/Oct/2007 Man Utd Middlesbrough HW 1.22 Betfair £30.54     27/Oct/2007 Reading Newcastle D 3.30 vcbet £30.54     27/Oct/2007 Sunderland Fulham AW 3.95 Betfair £30.54

[Matthew]

Date Home Away Prediction Odds Bookie Stake Correct Return 28/Oct/2007 Derby Everton D 3.25 BetFred £27.8     28/Oct/2007 Tottenham Blackburn AW 3.30 vcbet £27.8

[Ace123]

Hi Matthew,

Just a few questions re your model. Firstly i have to say im impressed but im fairly new to this whole thing and im extremely keen to build my own model.

The database you've created - how have you created this? is it in Excel/VBA or Access or another tool? And where are you collecting the data from?

What kind of predictive model have you built for the 3 outcomes? is it a linear regression model or logistic?

Finally, i see that you've developed some variables taken from a book. "System 9 Rating (A)" etc. Could you possibly tell me which book you got these from?

Hope to hear from you soon

Cheers

[Matthew]

I'm glad your taking this approach to predicting sports, it's applicable to
any sport really, you just have to get the data together.

I created my dataset exclusively in Excel using VBA to manipulate the data
and create the numerous variables you see on Page 1. The raw data if
you will is basically results info. Stuff like shots and corners are needed
for the "system" variables which come from the book Profitable Football
Betting written by Paul Steele.
http://www.highstakes.co.uk/shop/product.php/2172/

KXEN creates polynomial models and applies a technique known as SRM.

Structured Risk Minimization (SRM)

The primary challenge for statisticians has been to build highly accurate models that are also reliable. This has been particularly difficult given the large number of variables in computer-generated data. Traditional statistics can only produce an accurate model with a few variables, so an expert is needed to reduce the number of variables before building a model. The more variables there are, the more difficult it is to build a reliable model. Only the expertise of the statistician determines the reliability of the model.
SRM, developed by Vladimir Vapnik, is a breakthrough in mathematics and statistics that for the first time makes it possible to automatically build reliable and accurate models. In contrast to traditional statistical models, SRM models become more accurate and are still reliable as the number of variables is increased. Model Accuracy and Reliability are determined by the data, not by the expert.

What sort of variables are you looking to create?
Which leagues are you concentrating on?

You say you use SAS quite a lot? What do you do?

[Ace123]

I'm probably just looking to build a model based on the English Premier League. I think the variables that you've used seem pretty good so i might follow something similar.

I work as a statistician. Hence i know SAS very well. i know Excel/VBA quite well too and its a good way of cutting data but it would be sooooo much easier using SAS lol.

i take it that you built a polychotomous model ie a 3-way logistic regression model??

with regards to my other thread.......ive looked everywhere for it but no luck lol.

[Matthew]

@Ace123: Where have you looked thus far?

[Ace123]

ive looked at torrentspy.com and a few other websites but cant seem to find sas anywhere

any ideas??