Power Rankings (Human .v. Human games)

matthew8

Just throw up a system that gives me the Panthers. That's all I want. I don't want to be stuck with the Raiders and *shudder* Stanford Routt.

etloveslsd

empacher wrote: »

Also should it not be based exclusively on season 3, with anyone who has dropped out being given the latter picks?

If you use all 3 seasons, players like Paully, Steve-o, McG and daflood, should they all comeback would not get better teams as they would have play off records.

Say there are 16 teams with play off records, the will be the last 16 human picks. So the players with no play off records will have the earlier picks.

I think it's the fairest way, but that's my opinion.

helimachoptor

Danger_dave1 wrote: »

well the myself and the bengals would be 2 teams that would be out position and get teams better than are level within the BFL would warrant imo. Plus Santry not being last to pick even tho he's won every superbowl?

There's always statistical anomalies, also we can just make Santry the last pick!

For me I dont see a huge difference tbh, i havent won and play off games despite always making it to the play offs.

This isnt easy, also need to take into account newer players.

empacher

I'm just securing the pats for myself :cool:

etloveslsd

helimachoptor wrote: »

There's always statistical anomalies, also we can just make Santry the last pick!

For me I dont see a huge difference tbh, i havent won and play off games despite always making it to the play offs.

This isnt easy, also need to take into account newer players.

It does, nerd can still get into the play offs this year, so will matthew, the seahawks can also still make it.

ibmax

etloveslsd wrote: »

empacher wrote: »

I'd be in favor of the play off teams being the last 12 picks.

I think Ibmax means play off record throughout madden 12, so all 3 seasons.

I do... So whoever has the most playoff wins gets the later pick. Simples.

No farting about trying to figure out who beat who, who was on holidays and when, when someone joined and left and so on...

Everyone has opinions on who the best players are. These opinions are based on who's made the playoffs/made it to the superbowl/won the superbowl/etc and not who beat a CPU team by 80 points 3 weeks in a row!!! Basing ranking on anything other than playoffs means you end up with results where Santry is only the 4th best, which of course goes against everyone's opinions!!! Simples... again

ibmax

matthew8 wrote: »

Just throw up a system that gives me the Panthers. That's all I want. I don't want to be stuck with the Raiders and *shudder* Stanford Routt.

Stanford Routt is a Chief!!!

matthew8

ibmax wrote: »

Stanford Routt is a Chief!!!

Thank god, but they still suck. Just added the Cowboys to my list of approved teams (among realistic ones, 3 teams long) for next year, albeit with a change of scheme.

etloveslsd

helimachoptor wrote: »

For me I dont see a huge difference tbh, i havent won and play off games despite always making it to the play offs.

Well that would work in your favour then? Of all the teams with play off records, you would have one of the earlier picks!!!

helimachoptor

etloveslsd wrote: »

Well that would work in your favour then? Of all the teams with play off records, you would have one of the earlier picks!!!

ok on second thought its a greaty idea!!
i more meant i cant see me getting a huge advantage and ill havr my eye on 2 lesser teams anyway

ibmax

helimachoptor wrote: »

ok on second thought its a greaty idea!!
i more meant i cant see me getting a huge advantage and ill havr my eye on 2 lesser teams anyway

The way things are going all the "lesser" teams are all gonna be taken so I'll just sit back and wait for the packers

BigBadRob83

Was just putting this up as a discussion point, and yes Dave do enjoy statistical analysis, though not the compilation. Season 1 would be a nightmare to compile going through the old forum, that was main reason only did Seasons 2& 3. Unless people want to submt their own results for season 1, then can add that in.

Just a couple of points on the suggestions.
If its based on playoff wins, it would push the guys who have been here the longest down furthest as they would have more opportunities at playoffs.
Before we decide on a system, need to know what the effect is we're trying to achieve. Best teams to worst players or what?

The idea being using Human-v-Human results was to avoid some of the skewing of results that happens by using regular season overall record (which decides playoff results) which can be heavily influenced by how many CPU games a team plays and who is in their division.

Based on the data I've got, could work out the strength of schedule for each player which would give a weighting towards each win/loss which might be a better system? (Could loosely be described as the Berty effect).

Anyway, at the end of the day it's about what kind of system we want and what's the fairest, wherever your own team ends up within that

its less of a big deal for this because if theres no fantasy draft we can change teams every season

adrian522

Miracle Powerful Terminology wrote: »

its less of a big deal for this because if theres no fantasy draft we can change teams every season

Id have thought we would keep our team and try and improve through drafts and trades over 3 or 4 seasons

adrian522

adrian522 wrote: »

Id have thought we would keep our team and try and improve through drafts and trades over 3 or 4 seasons

thats a fair idea is there definitely an online franchise though

Danger_dave1

BigBadRob83 wrote: »

Before we decide on a system, need to know what the effect is we're trying to achieve. Best teams to worst players or what?

The idea being using Human-v-Human results was to avoid some of the skewing of results that happens by using regular season overall record (which decides playoff results) which can be heavily influenced by how many CPU games a team plays and who is in their division.

Based on the data I've got, could work out the strength of schedule for each player which would give a weighting towards each win/loss which might be a better system? (Could loosely be described as the Berty effect).

Anyway, at the end of the day it's about what kind of system we want and what's the fairest, wherever your own team ends up within that

Well said rob, makes sense about season 1, would be head wrecking. I think we should strive for the weakest players get to pick the stronger teams in my opinion. Realistically we'll prob only have 22/24 total start next season which means everyone is going to get a strong ish team and within 2 season will be able to compete for the SB

Anyone who decides to join the BFL and have never played before i would assume would go to the back of the line if they joined before the draft . Thoughts people ?

Could win's in the post season also be weighted ? like with the strength of schedule you talking about

---

Side note : Madden 13 if we play 5 days a week from the start we should get 4 seasons in. I presume everyone will be happy to go with that ?

linebacker52

adrian522 wrote: »

Id have thought we would keep our team and try and improve through drafts and trades over 3 or 4 seasons

i agree with adrain on this if there is only going to be 20-22 of us playing next season it wont matter as we would all get a fairly strong team anyway. if people pick the team they support eg titains bears loins there should be strong teams left for anyone that whated them.

BigBadRob83

Danger_dave1 wrote: »

Anyone who decides to join the BFL and have never played before i would assume would go to the back of the line if they joined before the draft . Thoughts people ?

Would agree with that alright

Danger_dave1 wrote: »

Could win's in the post season also be weighted ? like with the strength of schedule you talking about

Can do that.
Will do up a draft and see what people think, will attach the Excel sheet so everyone can see how it is worked out.
If I increase the importance of post-season games round-by-round maybe.

Danger_dave1 wrote: »

Side note : Madden 13 if we play 5 days a week from the start we should get 4 seasons in. I presume everyone will be happy to go with that ?

Definitely happy with that, 4 seasons would be savage

helimachoptor

New players should be left to last.

However what do we do about returning players?
McG, Stevo_o, mikemac etc?

A game every 5 days is good, as the beers people were chatting about on Friday, it allows you to pick up a good rated rookie and progress him.
It also means that giving up picks may have bigger impacts if you trade for an older player etc.

BigBadRob83

Explanation first of how this is worked out. Tried to use a similar formula to ones I have seen used on some college football websites to rank teams, using your opponents winning % to determine to 'strength' of each win/loss.

Worsksheets:
RESULTS-Tracks all the results for seasons 2 & 3
Records (2-3)-Summarises each teams records in different types of games
Weighting-This is where the Weighted Power Rankings are calculated
Picture-Where everything goes into a pretty picture (below)

Only Human-v-Human game results are used for this.

Each game is given a 'weighting', to take into account importance (to give heavier weight to more important games)
Regular Season = 1
Wildcard = 2
Divisional = 3
Conference = 5
Super Bowl = 10

The 'value' of each Human win is calculated as follows:
Weighting * Opponents' Human Winning Percentage

The 'value' of each Human loss is calculated as follows:
1 * (100% - Opponents' Human Winning Percentage)
Removed weighting from losses, as would actually end up making someone who lost Super Bowl's record worse

The final Power Ranking number (Weighted Winning Ratio) is calculated as follows:
(Sum total of all Weighted Wins / Sum total of all Weighted Losses)*2
Multiplied by 2 so that the average player should equal 1.00, just to give a reference point


Anyway, think it gives a fairly accurate representation (not 100% accurate, but that's never going to happen). Might update it every couple of weeks as more results come in. Obviously, more results more accurate it becomes.

EDIT: Just to give an example of what this is calculating. Take the Falcons and Eagles (both 5-11 vs. Human opponents).
This ratio ranks the Eagles higher as their victories have come against better opponents (55.4% vs 36.4%)

helimachoptor

I get me a good pick according to this

But that means i'm $hit

Berty44

For 4 seasons to work you need to run 2 seasons x 4 day gameweeks and then 2 seasons x 5 day week gameweeks. That way you just about will get them in.

Rob, fantastic work on the spreadsheet.

helimachoptor

Berty44 wrote: »

For 4 seasons to work you need to run 2 seasons x 4 day gameweeks and then 2 seasons x 5 day week gameweeks. That way you just about will get them in.

Rob, fantastic work on the spreadsheet.

I'd be happy to sign up to that tbh.
However it may mean some games are simmed due to availability.

BigBadRob83

If I've made mistakes on any of your results, you can make changes on the below Google Spreadsheet (Just make the appropriate change and note it in columns I & J so can track it), and I will update it next time I post it up.

https://docs.google.com/spreadsheet/ccc?key=0AsO5karsgEN6dGttMEU5OExHRGQwbGY3TEVKRTJCbGc

Hopefully LM will be sorted soon.

Berty44 wrote: »

For 4 seasons to work you need to run 2 seasons x 4 day gameweeks and then 2 seasons x 5 day week gameweeks. That way you just about will get them in.

If that's what we'd have to do for 4 seasons, I'd be on board (pun very much intended)

matthew8

Berty44 wrote: »

For 4 seasons to work you need to run 2 seasons x 4 day gameweeks and then 2 seasons x 5 day week gameweeks. That way you just about will get them in.

Rob, fantastic work on the spreadsheet.

I think early advances will allow for 4 seasons anyway.

Danger_dave1

matthew8 wrote: »

I think early advances will allow for 4 seasons anyway.

One thing I would be against is early advancement. Id like if its a 4 day game week , it's a 4 day game week, I'm part of a different league and to have to organize 3 games and then be told were advancing and I have less time to organize my next game would be a struggle.

It would set out a clear calendar for drafts/Superbowls etc.

helimachoptor

I'm kinda 50/50 on this, Dave I know what you are saying and pat of me is fully in support of it (and also my GF) I could schedule all my PFL and boards games a couple of weeks in advance pretty easily, however if we have all games done within 2 days, we really should be advancing.

However there shouldnt be a constant posting of "when are we advancing etc" not looking at anyone in particular :P

Berty44

I would also agree with Dave on this and in fact we were chatting about it at the beer meetup. It really helps particularly when you are in multiple leagues to know your schedule in advance so you can figure out when to schedule games. Also real life needs to take precendce and I find its handy to schedule a couple of games for the same night thereby completely freeing up other nights for living in the real world !

In practice if you are running a 4 days schedule you will find that rarely, if ever all the games will all be played one full day early so the Comm will most likeley not being able to advance anyway. More likely is that guys have not / cannot arrange to play the game in the time and you have to decide to boot one or other or let it go to sim. So better to not even consider it, as Dave says make a specific schedule and stick to it so we all know when games are due and particularly drafts take place.

matthew8

If we're not going to have early advances I think every gameweek should be 4 days. 5 is very long to wait when you get your game played within the first 2 day nearly every time.

adrian522

Every 4 days or 2 games per week to keep the days consistent.

5 days is along time when half the games are CPU games anyway. It might be different if we had 32 players.