Jump to content
** April Poker League Result : 1st Like2Fish, 2nd McG, 3rd andybell666 **

A Goal Difference System


Recommended Posts

I don't think this has been done before, but it's something that I'm very interested in. I have a theory where goal difference considerations are somewhat under rated in pricing up a match - if you take a look at any league table you will see goal differences for a team which are way out of line with their league position/points per game etc...my rationale here is that these teams are currently on the positive/negative side of variance and this will be corrected in the future. So I'm going to be looking at backing teams with an unreasonably high goal difference for their position, and opposing teams with an unreasonably low goal difference for their position. Staking will be on this basis:- Maximum lay liability £100. Any successful back profit £100.

Link to comment
Share on other sites

Re: A Goal Difference System Friday's bets:- Back Heerenveen v Utrecht @ 2.18 Lay Vannes at home to Sedan @ 3.0 Back Reims v Boulogne @ 2.04 Back 1860 Munich v Frankfurt @ 2.48 Back Braunschweig v Babelsburg @ 1.59 Back Den Bosch v Eindhoven @ 1.83 Lay Volendam v Go Ahead @ 4.3 Lay Gornik Zabrze v Lechia Gdansk @ 2.52 Lay Unirea Urziceni v Gloria Bistrita @ 3.4 Lay Barcelona II v Recreativo @ 3.0

Link to comment
Share on other sites

Re: A Goal Difference System Good luck sounds interesting just thought of something, could look at home and away league tables, maybe a team has a good goal difference and are on the up say away from home but just been a little unlucky, but it could be that the team is inconsistent have you back tested?

Link to comment
Share on other sites

Re: A Goal Difference System

Good luck sounds interesting just thought of something, could look at home and away league tables, maybe a team has a good goal difference and are on the up say away from home but just been a little unlucky, but it could be that the team is inconsistent have you back tested?
Thanks! That's something I've also thought about and something I may well bring in with time - I've noticed there are some teams (for example Crystal Palace) with an atrocious goal difference away from home, and would definitely be opposed in this situation. I haven't back tested at all but I'm only trialling this on paper at the moment. I won't be committing any money to it until at least 200-300 results.
Link to comment
Share on other sites

Re: A Goal Difference System

Thanks! That's something I've also thought about and something I may well bring in with time - I've noticed there are some teams (for example Crystal Palace) with an atrocious goal difference away from home, and would definitely be opposed in this situation. I haven't back tested at all but I'm only trialling this on paper at the moment. I won't be committing any money to it until at least 200-300 results.
If I may make a suggestion. Possibily a formula like this would help: home teams: Goal difference at home + 1/2*goal difference away. Away teams: Goal difference away + 1/2*goal difference at home. Of course the factor 1/2 could be changed to optimise your system. Good luck with this one.:hope
Link to comment
Share on other sites

Re: A Goal Difference System

If I may make a suggestion. Possibily a formula like this would help: home teams: Goal difference at home + 1/2*goal difference away. Away teams: Goal difference away + 1/2*goal difference at home. Of course the factor 1/2 could be changed to optimise your system. Good luck with this one.:hope
Thanks for the good wishes everyone! Nightfire, I like this. However the only problem at the moment is I dont have time to calculate these, although I really like the idea. Whereas I can take a cursory look at the tables and pick the teams out quickly at the moment. Having said this, if it works well I will definitely implement something like this too so we can see which variant works better.
Link to comment
Share on other sites

Re: A Goal Difference System There's an idea similar to yours, so you might find it interesting to read: over there at Understanding Uncertainty they tried to evaluate home and away "strength" and "weakness" depending on number of goals scored in comparison with average number of goals in that league: http://plus.maths.org/content/os/issue52/risk/index I modified somewhat that strategy and tested it previous year, but results were far from satisfying, so I eventually gave it up; I described it here: http://forum.punterslounge.com/f23/what-do-i-need-know-use-poisson-112355/ Hope you'll have more success with your approach! :hope

Link to comment
Share on other sites

Re: A Goal Difference System

Care to explain this in English?:unsure
Even this late in the season goals differences are not truly representative of team performance. Goals scored against differing opposition, at different locations should not all be treated equally. The transitive relationships in football carry certain implications. If team A plays team B, and team B plays team C, then the match between team A and team B is transitively connected to the match between team B and team C, since team B played in both. By exploiting these transitive relationships you can get a more accurate picture of the true weight of goals. Iterative processes maximize these relationships effectively. For football I use two models to produce match odds. I have a model that iterates over current season data and produces converged ratings of different teams. These ratings have been designed such that If I take team A's rating, and divide it by team B's rating, I will get a number that describes the ratio of team A's to team B's expected goals. E.g. Team A has a rating of 1700, team B has a rating of 1000, then team A will be expected to score 1.7 times as many goals as team B. That alone is useful, but doesn't tell us how many goals there will be in total, only how they will be broken down. My second model deals with the issue of total goals. Through an iterative process each team receives a converged pair of values that when combined with the values for another team describes the expected total goals in that game. Once we have that it is easy to use the other ratings to produce expected goals for each team, and then a suitable distribution (e.g. Poisson) can create match odds from these expectations. Anyway, the point really is that transitivity means that the result of every single game in the league has ramifications not just for the teams that played, but for all teams in the league, in terms of accurately predicting their scoring.
Link to comment
Share on other sites

Re: A Goal Difference System

...my rationale here is that these teams are currently on the positive/negative side of variance and this will be corrected in the future. So I'm going to be looking at backing teams with an unreasonably high goal difference for their position, and opposing teams with an unreasonably low goal difference for their position.
Wait a minute all "rocket scientists"!:) The initial idea is much simpler (see quote)! I would strongly suggest Sir ratholer to ignore all this sophisticated analysis and go forward with his initial concept - all things tend to go back to the mean. Actually I do believe that there might be some hidden advantages in this simplified approach. Why just not to test it for some decent number of matches.
Link to comment
Share on other sites

Re: A Goal Difference System Yes, that's definitely what I intend to do Mulkis. Just keep it simple at the moment and see where it takes us... Last night's matches:- Back Heerenveen v Utrecht @ 2.18 = +£100 Lay Vannes at home to Sedan @ 3.0 = +£50 Back Reims v Boulogne @ 2.04 = +£100 Back 1860 Munich v Frankfurt @ 2.48 = -£67.57 Back Braunschweig v Babelsburg @ 1.59 = -£169.49 Back Den Bosch v Eindhoven @ 1.83 = +£100 Lay Volendam v Go Ahead @ 4.3 = +£30.30 Lay Gornik Zabrze v Lechia Gdansk @ 2.52 = +£65.79 Lay Unirea Urziceni v Gloria Bistrita @ 3.4 = -£100 Lay Barcelona II v Recreativo @ 3.0 = +£50 7 winning bets, 3 losing bets. Overall profit so far £159.03. ROI Calculations to be added after this weekends matches...

Link to comment
Share on other sites

Re: A Goal Difference System Today's matches:- Lay West Brom v Chelsea Lay Scunthorpe v Crystal Palace Lay Exeter v Leyton Orient Lay Hartlepool v Sheff Wed Back Southampton v Bristol Rovers Back Crewe v Lincoln Back Shrewsbury v Stockport Lay Port Vale v Torquay Lay Wycombe v Northampton Back Grimsby v Bath Back Darlington v Hayes & Yeading (and lay Hayes and Yeading in a seperate bet) Back Luton v Mansfield Back Monaco v Nice Back Auxerre v Toulouse Back Stuttgart v Koln Lay Hannover v Hamburg Lay St Pauli v Wolfsburg Lay Werder Bremen v Schalke Back Getafe v Sevilla

Link to comment
Share on other sites

Re: A Goal Difference System

Today's matches:- Lay West Brom v Chelsea Lay Scunthorpe v Crystal Palace Lay Exeter v Leyton Orient Lay Hartlepool v Sheff Wed Back Southampton v Bristol Rovers Back Crewe v Lincoln Back Shrewsbury v Stockport Lay Port Vale v Torquay Lay Wycombe v Northampton Back Grimsby v Bath Back Darlington v Hayes & Yeading (and lay Hayes and Yeading in a seperate bet) Back Luton v Mansfield Back Monaco v Nice Back Auxerre v Toulouse Back Stuttgart v Koln Lay Hannover v Hamburg Lay St Pauli v Wolfsburg Lay Werder Bremen v Schalke Back Getafe v Sevilla
A slight problem with these picks. I completely forgot to check the prices so can't update the p&l :wall For interest only, 12 bets were winners and 8 bets were losers.
Link to comment
Share on other sites

Re: A Goal Difference System Tuesday's Results:- Lay York - L Lay Hayes and heading - L Lay wacker burghausen - W Back livingston (and lay Ayr) - both W Lay bellinzona - W Today's matches:- Lay Tottenham v Arsenal @ 3.0 Lay AAB v Randers @ 2.34 Lay Gornik Zabrze v Jagellonia @ 6 Lay St Gallen v Grasshopper @ 5.5

Link to comment
Share on other sites

Re: A Goal Difference System

A slight problem with these picks. I completely forgot to check the prices so can't update the p&l :wall For interest only, 12 bets were winners and 8 bets were losers.
You could fill the gap using this (BF, BQ lay prices at the bottom): http://www.oddsportal.com/soccer/england/premier-league-2010-2011/west-brom-chelsea-337047/
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...