Predicting The English Premier League-using stats

[Matthew]

Date Home Away Prediction Odds Bookie Stake Correct Return 18/08/2007 Tottenham Derby D 4.4 Betfair £34.13 0 18/08/2007 Newcastle Aston Villa AW 4.5 Betfair £34.13 0 18/08/2007 Portsmouth Bolton AW 4.4 Betfair £34.13 0 18/08/2007 Reading Everton D 3.35 Betfair £34.13 0

[Matthew]

		Bookmaker			Computer			Difference
Newcastle	Aston Villa	51%	30%	24%	32%	23%	50%	-19%	-7%	26%
Portsmouth	Bolton	53%	30%	25%	42%	9%	56%	-10%	-21%	31%
Reading	Everton	40%	31%	35%	34%	51%	21%	-6%	20%	-14%
Tottenham	Derby	67%	25%	14%	57%	44%	4%	-9%	19%	-10%

[happygooner]

Cheers for posting that extra info plopplop. Oh well, not long until the talking stops and we find out for real.

[happygooner]

Fair play to you with Boro plopplop (Kamara cracks me up - talk about excitable haha!!!) but jeeezzzz how much am I praying for a Villa win now!!

Out of interest, and should the worst happen, how long are you planning to give this before re-evaluating?

[Matthew]

This is turning into a fricking disaster isn't it.

Believe me - nobody is more disappointed than I am.

Not only is it fast turning into a LAY thread but it also flys in the face of all my hard work.

Looking at last year it appears that I am backing at way higher odds than I did last year and I don't think the fixtures have helped. But that is far from an excuse. I could never have forseen a 0 and 10 start.

It leaves me particularly dispondant.

But then it is only weekend #2 of the premiership and there is no way I can quit this early - despite how embarassing it is becoming.

You and I both know that the weekend I stop will be the first weekend i bag all 3/4/5, whatever the number is.

All in all I guess this is the danger of this type of system that relies on winning more often than not as opposed to trying to get every result right.

[Matthew]

but of course an AV win saves the day!!!

[happygooner]

Originally Posted by plopplop

but of course an AV win saves the day!!!

Haha, indeed it does......please Villa. They were playing well....

I guess you need to play it by ear really. See what the system comes up with for the next set of games and then evaluate whether you really see it happening.

I'll keep monitoring with interest as apart from you there's no one who wants to see this ship turned around more.

[trollerboy]

@plopplopp

a losing streak of 10 or 11 bets is not that incredible, especially given the high odds of your picks.

i´ve done some betting and in 860 bets i had a strike rate of approx. 60%. average odds were at about 1.85-1.9. losing streaks of 8 bets were very common.

so with your odds being even higher, it´s not too surprising to get this losing streak. the bad thing is just that you get it at the start. but it shouldn´t be taken too seriously.

so get on with it!

[2hcb]

chinup plopplop, your system isn't designed for every bet to win, you might hit 2/4 next time and you've wiped this week out already!

[SecondSight]

Ploppplop

I think Bolton situation has changed alot this season and you are probably using some past data that is rating them highly.

If you had laid Newcastle, Derby and Everton you would have won and it would have shown these bets as value on your system.

Your system will probably work better after the first few games of the season are out of the way.

You need the new Boro Badge!

[chilii]

Plopplop,I have to ask you this,please don't think I am undermining your work.

Your system is saying that probability of draw in Reading-Everton match is 51%.How can this be true in regular league game(except in Italy)?Also for Portsmuth-Bolton only 9% chance for draw?Did the old system showed similar probabilities?
GL.

[Matthew]

Apologies to all. Very late with these:

Date Home Away Prediction Odds Bookie Stake Correct Return 19/08/2007 Blackburn Arsenal D 3.35 Betfair £28.67 0 19/08/2007 Chelsea Liverpool D 3.25 Betfair £28.67 0 19/08/2007 Man City Man Utd D 3.45 Betfair £28.67 0

[happygooner]

Originally Posted by plopplop

Apologies to all. Very late with these:

Err....Man U already kicked off haha!! Best of luck anyway.

[merlin]

.... I think that the root cause of such a poor start PP is probably down to using old data such as previous seasons.Circumstances change from one season to another as we all know, new managers and players come and go and along with early season form or indeed in many cases, lack of this all upsets the course of events.If this is artificial intelligence being used to produce your predictions, then it will be way off the mark.Iv'e been there,done it, and got the teeshirt!I found out the hard way, no matter how you input data, the answers churned out are always disappointing for the amount of effort put in.The concept always looks promising, but alas the results are never quite up to what you expect.I urge you to have a serious rethink!

[Hodgey]

FT

Blackburn R.

1 - 1

Arsenal

Well done mate. Must be a relief. Lets hope it's the start of the turnaround.

[happygooner]

Well who knows where you're going from here but congrats on a good winning day anyway. I know my bank looks healthier for it.

Cheers again

[Hodgey]

Just like buses - nothing for ages then 2 come along at once!

Well done mate. Things looking a bit brighter already.

[trollerboy]

so far plopplop has given 14 picks.

average odds: 3.57

even with an ROI of 120% the strike rate would only be at 33-34 %.

generally speaking , the strike rate will roughly be in the region of around 30% (somewhere between 25-35%).

with a strike rate low as this, losing streaks of 12 losing bets are not really uncommon.

so, the bad start just means nothing. it´s just very normal. one has adjust the stake size to the low strike rate.

[Notreve1977]

Your selection method appears to be showing general balue in Draws i beleive peoples tendancy to pick a winner rather than the draw is the cause of this

[Matthew]

Where to start.................

Thanks to London_calling, 2hcb, trollerboy, Hodgey and happygooner for your best wishes and good luck messages - they are appreciated.


To answer some more specific questions:

Originally Posted by algorithmbetting

i find the wide divergence between bookies odds and your odds interesting. my odds too are often very different than bookies though my system is not as refined as yours i think.

Originally Posted by algorithmbetting

when i initially noted the large difference in odds i thought that my model must be wrong/stupid but then why should estimates of % win/loss probability vary by only +/-10% for instance. given west ham can do the double over arsenal and man u in one season and yet their odds pre match no way refelcted this. the fact that market odds across bookies/exchange rarely vary by more than 1/2% is purely down to an active market rather than confirming a true reflection of the fair odds.

not sure if that made sense but be interested in your view of the wide range in odds

I think this is undoubtedly down the lack of information available with which to refine your odds - if you are finding wild variations come the season end then I'd suggest there might be something awry in your methods. I know for sure that my %ages become tighter as the season moves on.

Originally Posted by SecondSight

You need the new Boro Badge!

I dislike the new Boro badge.

Originally Posted by chilii

Plopplop,I have to ask you this,please don't think I am undermining your work.

Originally Posted by chilii

Your system is saying that probability of draw in Reading-Everton match is 51%.How can this be true in regular league game(except in Italy)?Also for Portsmuth-Bolton only 9% chance for draw?Did the old system showed similar probabilities?

GL.

Why can't it be true that a game has a high chance of being a draw? Surely you must agree that 2 teams playing each other with solid defences and lack of scoring prowess are more likely to draw than a strong offensive and defensive team against a team that are quite simply hopeless at both.

My old system produced %ages just as varied as this one. Albeit they did settle down as the season progressed - but again - it depended on the merits of the teams playing the game, who was at home, who was away.

Originally Posted by merlin

.... I think that the root cause of such a poor start PP is probably down to using old data such as previous seasons. Circumstances change from one season to another as we all know, new managers and players come and go and along with early season form or indeed in many cases, lack of this all upsets the course of events.If this is artificial intelligence being used to produce your predictions, then it will be way off the mark.Iv'e been there,done it, and got the teeshirt!I found out the hard way, no matter how you input data, the answers churned out are always disappointing for the amount of effort put in.The concept always looks promising, but alas the results are never quite up to what you expect.I urge you to have a serious rethink!

I heed your warning - and starting from the off was a gamble in itself. The reason I choose to do so is because with this version of the model last year I would have made 9.9% yield. I've backed at higher odds this month than I would have last year with this model, with the sample being so small to date and the fixtures not really going in my favour then it's no surprise I'm behind. It's probably just a little more embarrassing with it being in public like this and I also have a minority of the 22 people in my syndicate on my back.

I'd like you to PM me, or even reply with your experience (i.e. having the t-shirt I may have to have come the end of this season) and also what in particular you think I should rethink.

Do you think my whole line of attack is wrong?
Am I being too aggressive?

I'm always open to listening to the experienced people from PL that take time out to post in my thread.


AT CHUFFING LAST!!!!!!!!!!!!!

Date Home Away Prediction Odds Bookie Stake Correct Return 19/08/2007 Blackburn Arsenal D 3.35 Betfair £28.67 Y £92.67 19/08/2007 Chelsea Liverpool D 3.25 Betfair £28.67 Y £89.95 19/08/2007 Man City Man Utd D 3.45 Betfair £28.67 N

System Stats:

ALL HW D AW No of Bets: 14 0 9 5           Staked: £519.37 £0 £329.22 £190.15 Returned: £182.62 £0 £182.62 £0 P+L: £-336.75 £0 £-146.6 £-190.15 Strike Rate: 14.3% 0% 22.2% 0% Yield: -64.8% 0% -44.5% -100%           Avg Back Odds: 3.57 0 3.62 3.48 Avg Win Odds: 3.3 0 3.3 0           Each Person: £35.36