Goal Supremacy

[barker8]

Hey guys,

First time post so bear with me! Trying to get my head around Goal Supremacy and working out my own odds but they seem a little 'off'. I'm sure this is normal until I figure out a way that works for me, but initial calculations are throwing me off a bit. I'm looking at a site that has a post on compiling odds (probably not best to post that here, or can I?).

I worked out some calculations for tonight's game in the Championship. So I have the following data:

(Goals For Leeds average 0.9666 + Boro Against average 0.5517) / 2 = 0.7592
(Goals For Boro average 1.3448 + Leeds Against average 1.133) / 2 = 1.2389

So that gives a Total Goals of 1.9981.
And a Supremacy of (0.7592 - 1.2389 + 0.4) -0.08

Compiler Odds

Home 3.07
Draw 3.24
Away 2.73

Under 2.5 1.47
Over 2.5 3.09

Actual Betfair Odds

Home 4.10
Draw 3.35
Away 2.16

Under 2.5 1.61
Over 2.5 2.62

 

Reason for the compiler odds being so far off?

 

Thanks guys, appreciate any replies.

[Sir Puntalot]

 to PL Barker8.  

[barker8]

Thanks Paul  Ok for me to post a link to that external site or is that a no no here? Just might help explain what I unsuccessfully tried to explain above!

[harry_rag]

Your goals and supremacy numbers are probably unrealistic. The spreads imply total goals are expected to be around 2.2 or 2.25 with Boro supremacy 0.4 or 0.45. Maybe those numbers would give you closer to the market odds. If you really trust your numbers, then take the value you've found (but if it was that easy, everyone would be doing it)!

[harry_rag]

Just to add, you'll never get true numbers just by crunching the averages for each team this season, you'll need to be far more sophisticated than that to beat the bookies but, if you are going to do that, your figures in this case have been badly skewed by the unusually low number of goals Boro have conceded at home. As they're away, you'd arguably be better (but not that much) focusing on respective home/away totals for each team.

[allen29]

Because you are using a simple formula and goals are not the best measure of team quality. They are rare events and subject to chance. Also, a high scoring game can skew a team's average. 

You may want to read about Expected Goals. Here is a good article.

Edited February 15, 2016 by allen29

[barker8]

Sorry guys, been away for the past few days so didn't get a chance to reply. I should have probably stated that I'm just looking for a bit of "fun" with a system, something to remove the guesswork but also something I can play around with (and not take it too seriously!). For example, I've been having a look at the setup from this site (http://www.jollyodds.com/compiling-football-odds/). I can just about get my head around it but wouldn't mind tweaking it a little bit. Let's say I do the basic calculation (as above in the Boro game) but look at the team news and notice that some key players are missing from the lineup. What's the best way of reflecting this using that spreadsheet in the link? If I just adjust the supremacy value, it only affects the match odds (has no effect on overs).

[harry_rag]

If the significant absence was in the underdog's defence, for example, you could increase the total goals expectation and the supremacy by the same amount, e.g. add 0.2 onto both so that 1.5/1.1 became 1.7/1.1. That would impact match and overs odds.

[barker8]

Thanks Harry. So I've inputted the below data and it's thrown out the figures in the screenshot, no other adjustments made.

Home Team - 22 Games Played, 24 Goals For, 28 Goals Against

Away Team - 22 Games Played, 30 Goals For, 24 Goals Against

 

Actual Odds:

Home 2.16

Draw 3.40

Away 4.00

Over 2.5 2.46

Under 2.5 1.67

 

 

Overs are not too far off what the market is saying but the match odds are, once again, way off. The formula obviously hasn't taken anything else into account (injuries, form etc). The away team have been in poor form, not won in six games. So I increase the total goals expectation (2.409+0.2) and supremacy (0.173+0.2). This gives me the below:

 

 

Probably a dumb question but why does this bring down the home odds, but also significantly bringing the overs odds down as well?

[harry_rag]

Because you've increased the total goals expectation from around 2.41 to 2.61, simple as that. As the calculator seems to be pretty much be using poisson to calculate the odds, it's underestimating the chances of both teams scoring the same number of goals (i.e. the draw).

As has been stated, just using the data you're using will never get you that close to the "true" odds but it's fine if you're just dipping your toe in the waters or are doing it for fun. Even if you used numbers based on 50% of your calculation and 50% of the average home/away goals for the whole league it would probably get you closer to the market odds.

[barker8]

It's actually using bivariate poisson, not that it makes a whole lot of sense to me! Is there a way of adjusting the model to stop it from underestimating the draw?

 

I understand that what I'm doing is still way off, but I wouldn't mind a little bit of reassurance that I'm not way off the mark  Like you say, it's just a bit of poorly calculated fun!

[barker8]

Sorry to bump this but still wondering if there's any way to adjust this to prevent it from underestimating the draw?

[MaxRedd]

I read a good post about goal supremacy rating here http://betbubbles.com/goals-supremacy-rating/

i hope punterlounge menbers will like:-)

[paparainbow]

considering your initial input,

expected home 0.90

expected away 1.35

 

 

you have to revisit your mathematical model.

 

 

Edited July 17, 2017 by paparainbow

[Powerade]

hello guys, does anyone still have the file and could make it available for download?

 

Thank you very much