Coefficient Ranking

Milk not bear jizz

Milk not bear jizz

Grasser-In-Chief
Nov 12, 2013
28,188
10,177
113
Mengbilla
[table="width: 500"]
[tr]
[td]Manchester City[/td]
[td]100[/td]
[/tr]
[tr]
[td]Liverpool[/td]
[td]95[/td]
[/tr]
[tr]
[td]Chelsea[/td]
[td]90[/td]
[/tr]
[tr]
[td]Arsenal[/td]
[td]85[/td]
[/tr]
[tr]
[td]Everton[/td]
[td]80[/td]
[/tr]
[tr]
[td]Tottenham[/td]
[td]75[/td]
[/tr]
[tr]
[td]Manchester United[/td]
[td]70[/td]
[/tr]
[tr]
[td]Southampton[/td]
[td]65[/td]
[/tr]
[tr]
[td]Stoke City[/td]
[td]60[/td]
[/tr]
[tr]
[td]Newcastley United[/td]
[td]55[/td]
[/tr]
[tr]
[td]Crystal Palace[/td]
[td]50[/td]
[/tr]
[tr]
[td]Swansea City[/td]
[td]45[/td]
[/tr]
[tr]
[td]West Ham Utd[/td]
[td]40[/td]
[/tr]
[tr]
[td]Sunderland[/td]
[td]35[/td]
[/tr]
[tr]
[td]Aston Villa[/td]
[td]30[/td]
[/tr]
[tr]
[td]Hull City[/td]
[td]25[/td]
[/tr]
[tr]
[td]West Brom Albion[/td]
[td]20[/td]
[/tr]
[tr]
[td]Leicester City[/td]
[td]15[/td]
[/tr]
[tr]
[td]Burnley[/td]
[td]10[/td]
[/tr]
[tr]
[td]QPR[/td]
[td]5[/td]
[/tr]
[/table]


What is this? It is an experiment to try and take a pulse of the Premier League- find out how good teams CURRENTLY are by- how well they played- being awarded more points for beating other teams currently doing well. The most recent results will count more towards your ranking than results from longer ago- and beating bigger teams will earn you more points than beating smaller teams.


To start off with- I am seeding the coefficients starting with 100 for the top of last year's league down in steps of 5. I have devised a formula to hereafter approximate how good a team is based on who they beat and by how much. (more points for beating a big team- more points lost for losing to a small team relative to oneself).



Points will be awarded thus:
If team A beats team B- Team A will get 10% of Team B's coefficient added to their coefficient. (example Liverpool beat Man City their new coefficient would be 95 + 10 = 105)
If team A beats team B by 3 or more goals they get 20% of Team B's coefficient added to their coefficient. (example Liverpool beat Man City by 3 goals their new coefficient would be 95 + 20 = 115)

If the game is a draw- the team with the high coefficient going into the game will drop 20% of the gap between the two teams and the team with the lower coefficient going into the game will go up 20%.
(Example Liverpool draw with Man City. 5 point different in coefficient, so Liverpool would go up 1 pt to 96. Man City would drop to 99)



If Team A loses to Team B and Team A has a lower coefficient than Team B- team A will drop 10% of their coefficient. (example Liverpool lose to City- they will drop 9.5pts to 86.5)
If Team A loses to Team B by 3 or more goals and Team A has a lower coefficient than Team B- team B will drop 20% of their coefficient. (example Liverpool lose to City- they will drop 19pts to 76)

If Team A loses to Team B and Team A has a higher coefficient- Team A will incur the additional penalty of dropping 10% of the difference between the two teams. (If City lose to Liverpool- they would drop the standard 10% to 90- and then in addition drop another .5pts for losing to a team with a 5pt lower coefficient than themselves- so would end up with 89.5pts).

At the end of this season- each team will carry their coefficients over to the next season. If no team has a coefficient of 100 or greater- all teams will receive same % increase so that the highest ranked team has a 100 coefficient.

The three teams being promoted will receive seed coefficients up to 15, 10, and 5 respectively (capped out so as not to surpass any team that was in Premier League the season before).

I will round coefficients to nearest tenth.





We've a while before the season starts... and suggestions on how to make the coefficient system better? (I don't promise to listen to anyone).
 
You either really love maths or you have no idea how much work this will involve over the course of the season <ok>
Either way, I hope you're unemployed to find the time for this

Would be an interesting experiment though.
 
cini65 said:
Or we could just use the Premier league table?
Click to expand...

Where's the fun in that?!?!




This measures something different... this is more to measure how good a team CURRENTLY is. (will take several rounds to become accurate).

If a team is crap for 6 months and is good the last 6 months- they have no hope of winning the PL- even if they could be, for the last 6 months the best team in the league. This coefficient ranking is designed to give a pulse onhow good a team is currently- not how good they were over the course of a season.
 
Two girls said:
You either really love maths or you have no idea how much work this will involve over the course of the season <ok>
Either way, I hope you're unemployed to find the time for this

Would be an interesting experiment though.
Click to expand...

It would take ten-fifteen minutes per round of fixtures
 
Two girls said:
You either really love maths or you have no idea how much work this will involve over the course of the season <ok>
Either way, I hope you're unemployed to find the time for this

Would be an interesting experiment though.
Click to expand...



Took 10 minutes to write an Excel app to calculate it for me. I'd just need to copy the scores in and it spits the table out for me.
 
Gerrez said:
It would take ten-fifteen minutes per round of fixtures
Click to expand...

I write the program first... then it only takes as long as it takes to copy and paste the scores into my excel app. Shouldn't take more than a minute....


If I find an XML feed of all the scores- I can get it to recalculate for me without me having to do anything.
 
Interesting.
Be good to keep an average of a teams co-efficient as well so at the end you can compare against the proper league table and see how accurate it actually is.
 
Milk said:
Took 10 minutes to write an Excel app to calculate it for me. I'd just need to copy the scores in and it spits the table out for me.
Click to expand...

Have you written the equations with each of these different bonus factors as different paramaters so you can change them if necessary and instantly recalculate the coefficient time series? There might be some balancing needed.
 
astroturfnaut:6768785 said:
Milk said:
Took 10 minutes to write an Excel app to calculate it for me. I'd just need to copy the scores in and it spits the table out for me.
Click to expand...

Have you written the equations with each of these different bonus factors as different paramaters so you can change them if necessary and instantly recalculate the coefficient time series? There might be some balancing needed.
Click to expand...


Kind of... as is currently written need to put result check in correct column.

I will rewrite one weekend to automate everything for me. Perhaps make a web app so people can view results.
 
Fair play to you milk. **** the begrudgers I'll be tuning in each week and using your co efficient for my football bets.

If I lose a fortune theres not a place on earth you can hide (besides your house as I dont know your address).
 
Hash said:
Fair play to you milk. **** the begrudgers I'll be tuning in each week and using your co efficient for my football bets.

If I lose a fortune theres not a place on earth you can hide (besides your house as I dont know your address).
Click to expand...

America is his address.





P.S. Good work Milk, you're some crazy ****ing genius.
 
Milk said:
I write the program first... then it only takes as long as it takes to copy and paste the scores into my excel app. Shouldn't take more than a minute....


If I find an XML feed of all the scores- I can get it to recalculate for me without me having to do anything.
Click to expand...

Why are the scum top? <grr>
 
You must log in or register to reply here.
Top Bottom
Back