Coefficient Ranking

So... this is about who's currently best..

so here's my question.

city have 100 points

Qpr have 5.

so what happens if city do a utd and end up 7th and qpr are 8th for whatever reason. can QPR ever catch city in this system?

example:

city beat every lower team in the table and lose to every top 8 side.. QPr can randomly win and draw throughout the league. Can qpr ever close the gap form last season?

second question:

if utd are top <puke> but have 70 starting point and city are six points behind but have 100 points half way through would utd have better coefficient?

finally.... could you have a scenario where city win 12 in a row but most are v lower teams so they could be only seocnd best team?
 
[video=youtube;Tk6Mjg3_eq0]https://www.youtube.com/watch?v=Tk6Mjg3_eq0[/video]
 
Keep us updated, Milk. I'll be following the results of your experiment with interest.

Seems a promising thesis. As for improving it, I have no idea, really. Perhaps, a modifying factor for squad size and games played?

But why tinker with it now? Try it for a season and see how it works out. No doubt, during the season, any flaws in the system will become apparent, and you can then think about ways of addressing those shortcomings. <ok>
 
moreinjuredthanowen said:
So... this is about who's currently best..

so here's my question.

city have 100 points

Qpr have 5.

so what happens if city do a utd and end up 7th and qpr are 8th for whatever reason. can QPR ever catch city in this system?

example:

city beat every lower team in the table and lose to every top 8 side.. QPr can randomly win and draw throughout the league. Can qpr ever close the gap form last season?

second question:

if utd are top <puke> but have 70 starting point and city are six points behind but have 100 points half way through would utd have better coefficient?

finally.... could you have a scenario where city win 12 in a row but most are v lower teams so they could be only seocnd best team?
Click to expand...



QPR can overtake City.

If City lose three in a row against teams with an average of "10" for example.
Start: Coefficient 100
After Game 1: -10 -9 (for losing to team 90 pts lower) = New Coefficient 81
After Game 2: -8.1 -7.1(for losing to team 71 pts lower) = New Coefficient 65.8
After Game 3: -6.58 -5.58 = New Coefficient 53.64

So- City, if they lose against 3 absolutely Shiite teams in a row- their coefficient will drop significantly.

Now take QPR - it they win against three top-ranked teams (say 95 average coefficient) by a three goal margain each time (so get the bonus 10%)
Start: 5
After Game One: +9.5 +9.5(for three goal winning margain) New Coefficient 5 + 19 = 24
After Game 2: 24 +19 = 43
After Game 3: 43 + 19 = 62



So- theoretically- if City lose against really crap teams for three games in a row. AND QPR absolutely demolish top ranked teams for three games in a row- yes QPR could have a higher coefficient than City in as few as three games. In reality though- it's unlikely for a team to drop quite so quickly (unlikely to play three absolute crap teams in a row... and lose). AND... QPR is unlikely to demolish three high-coefficient teams in a row by a 3 goal margain.

It would probably take 7 or 8 games of QPR doing well and City doing badly for QPR to overtake City's coefficient because they are starting out so far apart... Teams closer together will likely change places a lot more.

A mid table team would be able to overtake City easier- which is fine.. because, it is supposed to be a relative measure of how good a team has been recently... not necessarily represent who is the best over a course of a season.
 
CCC said:
Keep us updated, Milk. I'll be following the results of your experiment with interest.

Seems a promising thesis. As for improving it, I have no idea, really. Perhaps, a modifying factor for squad size and games played?

But why tinker with it now? Try it for a season and see how it works out. No doubt, during the season, any flaws in the system will become apparent, and you can then think about ways of addressing those shortcomings. <ok>
Click to expand...

Yeah... I figured I may need to play with the formulas somewhat during that first year. I'm not sure how I could objectively work in Squad Size into the equation.

I have wondered if I should take Home/Away into consideration. Perhaps give an extra few % boost for winning away- and an extra few % drop if you lose away. In event of a draw, perhaps giving a small bonus to the away team.


I'm also going to need to watch how well my increase vs decrease formulas are balanced... if everyone's coefficients will as a trend drift up or down over a season. When I have some spare time- I may plug last seasons results in and see what the coefficient spread looks like after a season.
 
Milk said:
QPR can overtake City.

If City lose three in a row against teams with an average of "10" for example.
Start: Coefficient 100
After Game 1: -10 -9 (for losing to team 90 pts lower) = New Coefficient 81
After Game 2: -8.1 -7.1(for losing to team 71 pts lower) = New Coefficient 65.8
After Game 3: -6.58 -5.58 = New Coefficient 53.64

So- City, if they lose against 3 absolutely Shiite teams in a row- their coefficient will drop significantly.

Now take QPR - it they win against three top-ranked teams (say 95 average coefficient) by a three goal margain each time (so get the bonus 10%)
Start: 5
After Game One: +9.5 +9.5(for three goal winning margain) New Coefficient 5 + 19 = 24
After Game 2: 24 +19 = 43
After Game 3: 43 + 19 = 62



So- theoretically- if City lose against really crap teams for three games in a row. AND QPR absolutely demolish top ranked teams for three games in a row- yes QPR could have a higher coefficient than City in as few as three games. In reality though- it's unlikely for a team to drop quite so quickly (unlikely to play three absolute crap teams in a row... and lose). AND... QPR is unlikely to demolish three high-coefficient teams in a row by a 3 goal margain.

It would probably take 7 or 8 games of QPR doing well and City doing badly for QPR to overtake City's coefficient because they are starting out so far apart... Teams closer together will likely change places a lot more.

A mid table team would be able to overtake City easier- which is fine.. because, it is supposed to be a relative measure of how good a team has been recently... not necessarily represent who is the best over a course of a season.
Click to expand...

so you think after what 3-5 games in the effect of the "old season" fades out. and then by 10 games really you should be on new season only?

Is there only a finite total of points in your system, ie the sum of starting points? eg. if say city trash a team on 5 points they can only get 20% of 5

in other words someone cannot really accumulate 50% of the points in the game... as it were. eg. if i play say QPR after city thrash them 5-0 they have a lower total so i can only take that lower proportion despite also thrashing them 5-0.
 
moreinjuredthanowen said:
so you think after what 3-5 games in the effect of the "old season" fades out. and then by 10 games really you should be on new season only?

Is there only a finite total of points in your system, ie the sum of starting points? eg. if say city trash a team on 5 points they can only get 20% of 5

in other words someone cannot really accumulate 50% of the points in the game... as it were. eg. if i play say QPR after city thrash them 5-0 they have a lower total so i can only take that lower proportion despite also thrashing them 5-0.
Click to expand...

All old results will be reflected in your coefficient- but the more recent a result the more it will affect your coefficient. I would say, that, yes, I believe the last 5 or 6 results impact your coefficient more than the games that came before it... I will need to tinker and run some samples through to see... I've not done any long-term testing of my figures yet.

You get 10% of the other teams coefficient added to your own for beating them.
You get 20% if you win by 3 or more goals.

So, yes, there is a bigger advantage to winning by a large margain... I don't think giving any extra reward for any other win 5+, etc really adds that much. Once you get 3+ goals, I think there is a psychological switch where teams just collapse- and that isn't really indicative of how much better a team really is.
 
You could just run it from the start of last season, or from the start of 2014 since newer fixtures dominate the result
 
moreinjuredthanowen said:
so you think after what 3-5 games in the effect of the "old season" fades out. and then by 10 games really you should be on new season only?

Is there only a finite total of points in your system, ie the sum of starting points? eg. if say city trash a team on 5 points they can only get 20% of 5

in other words someone cannot really accumulate 50% of the points in the game... as it were. eg. if i play say QPR after city thrash them 5-0 they have a lower total so i can only take that lower proportion despite also thrashing them 5-0.
Click to expand...

It was a bloody example to answer your negative question :p
 
astroturfnaut said:
You could just run it from the start of last season, or from the start of 2014 since newer fixtures dominate the result
Click to expand...


Good idea! I may do that... start it from the beginning of last season! I'd be curious to see how it ends up... we had a rather poor string of results at the end of the season- so I suspect we're several spots below 2nd...
 
Milk said:
Good idea! I may do that... start it from the beginning of last season! I'd be curious to see how it ends up... we had a rather poor string of results at the end of the season- so I suspect we're several spots below 2nd...
Click to expand...

Chelsea were also **** though

Sunderland will be right up there
 
You must log in or register to reply here.
Top Bottom
Back