+- CSNbbs (https://csnbbs.com)
+-- Forum: Active Boards (/forum-769.html)
+--- Forum: CUSAbbs (/forum-514.html)
+---- Forum: CUSA Conference Talk (/forum-439.html)
+---- Thread: Working on something and need input (/thread-791081.html)
Pages: 1 2
Working on something and need input - banker - 09-16-2016 10:57 AM
I am working on a ranking system (objective) for G5 programs and am developing a scoring system. Here's the rough start, but I would like some input.
Versus FCS
Home Win 1
Home Loss -3
Neutral Win 1.5
Neutral Loss -2
Away Win 2
Away Loss -1
Versus Unranked G5
Home Win 2
Home Loss -2
Neutral Win 2.5
Neutral Loss -1.5
Away Win 3
Away Loss -1
Versus Unranked P5
Home Win 3
Home Loss -1.5
Neutral Win 4
Neutral Loss -1.0
Away Win 5
Away Loss -0.5
Versus Ranked Team
Home Win 5
Home Loss 0
Neutral Win 6
Neutral Loss 0
Away Win 7
Away Loss 0
RE: Working on something and need input - Georgia_Power_Company - 09-16-2016 11:09 AM
(09-16-2016 10:57 AM)banker Wrote: I am working on a ranking system (objective) for G5 programs and am developing a scoring system. Here's the rough start, but I would like some input.
Versus FCS
Home Win 1
Home Loss -3
Neutral Win 1.5
Neutral Loss -2
Away Win 2
Away Loss -1
Versus Unranked G5
Home Win 2
Home Loss -2
Neutral Win 2.5
Neutral Loss -1.5
Away Win 3
Away Loss -1
Versus Unranked P5
Home Win 3
Home Loss -1.5
Neutral Win 4
Neutral Loss -1.0
Away Win 5
Away Loss -0.5
Versus Ranked Team
Home Win 5
Home Loss 0
Neutral Win 6
Neutral Loss 0
Away Win 7
Away Loss 0
I like it! I proposed something similar over on the Sun Belt Board which would better weight games vs FCS and Ranked teams than what most writers do.
(09-15-2016 11:59 AM)Georgia_Power_Company Wrote:(09-15-2016 10:28 AM)FoUTASportscaster Wrote: Crap rating that does not weigh for conference members. With every conference having more teams, even an FCS win is going to push them ahead. Quantity does not equal quality.
I agree completely.
To better grade the G5 conferences I would propose
1 point for FCS Win/-3 for a FCS loss
2 points for a G5 Win/-2 for a G5 loss
3 pints for a P5 win/-1 for a P5 loss
3 bonus points for a win over a ranked FBS/2 bonus points for a one score loss to ranked FBS/1 bonus point for a loss to a ranked FBS (unless team qualified for the 2 point bonus)
All conference games are 0/0
After adding up all of the points you divided them by the number of teams to get the conference average.
So using my formula after week 2 the G5 ranks:
1. American (1.92)
2. MW (1.08)
3. Sun Belt (0.45)
4. MAC (0.33)
5. CUSA (0.15)
RE: Working on something and need input - ThreeifbyLightning - 09-16-2016 11:14 AM
You lose objectivity with variations on performance based against ranked opponents. Rankings are inherently subjective; therefore, any "system" with subjective inputs automatically eliminates the viability of objectivity.
Also the point values are created with bias, which also undermines objectivity.
The problem here is that all rankings and systems (both human and computer generated) have inherent flaws because no one has determined a method to create true objective data for the purposes ranking teams when you have a 130+ teams and only 12 games. The human polls are obviously anything but objective and there isn't enough data to create truly objective system otherwise.
But good luck.
RE: Working on something and need input - TribeNiner - 09-16-2016 11:19 AM
I think that you may need to break it down further into RPI levels (1-25 is worth X, 26-50 is worth X, etc.). I'd be more impressed by a road (or home) win against NDSU than a home win against Iowa State or UVA.
RE: Working on something and need input - Georgia_Power_Company - 09-16-2016 11:24 AM
(09-16-2016 11:14 AM)ThreeifbyLightning Wrote: You lose objectivity with variations on performance based against ranked opponents. Rankings are inherently subjective; therefore, any "system" with subjective inputs automatically eliminates the viability of objectivity.
Also the point values are created with bias, which also undermines objectivity.
The problem here is that all rankings and systems (both human and computer generated) have inherent flaws because no one has determined a method to create true objective data for the purposes ranking teams when you have a 130+ teams and only 12 games. The human polls are obviously anything but objective and there isn't enough data to create truly objective system otherwise.
But good luck.
No system is perfect but teams are ranked for a reason even if that reason is flawed. The entire reason we rank teams and in this case conferences is to perpetuate a sense of order out of chaos and therefore it is only natural to weight wins or loses based on the perceived value of the team we played.
RE: Working on something and need input - ThreeifbyLightning - 09-16-2016 11:33 AM
(09-16-2016 11:24 AM)Georgia_Power_Company Wrote:(09-16-2016 11:14 AM)ThreeifbyLightning Wrote: You lose objectivity with variations on performance based against ranked opponents. Rankings are inherently subjective; therefore, any "system" with subjective inputs automatically eliminates the viability of objectivity.
Also the point values are created with bias, which also undermines objectivity.
The problem here is that all rankings and systems (both human and computer generated) have inherent flaws because no one has determined a method to create true objective data for the purposes ranking teams when you have a 130+ teams and only 12 games. The human polls are obviously anything but objective and there isn't enough data to create truly objective system otherwise.
But good luck.
No system is perfect but teams are ranked for a reason even if that reason is flawed. The entire reason we rank teams and in this case conferences is to perpetuate a sense of order out of chaos and therefore it is only natural to weight wins or loses based on the perceived value of the team we played.
Again, perceived value is subjective.
Feel free to come up with any ranking desired, but it doesn't make it objective.
RE: Working on something and need input - pilot172000 - 09-16-2016 12:39 PM
Now, does the team have to be ranked when you play them or can they be like a week or two later?
RE: Working on something and need input - goherd24herdfans - 09-16-2016 12:54 PM
(09-16-2016 10:57 AM)banker Wrote: I am working on a ranking system (objective) for G5 programs and am developing a scoring system. Here's the rough start, but I would like some input.
Versus FCS
Home Win 1
Home Loss -3
Neutral Win 1.5
Neutral Loss -2
Away Win 2
Away Loss -1
Versus Unranked G5
Home Win 2
Home Loss -2
Neutral Win 2.5
Neutral Loss -1.5
Away Win 3
Away Loss -1
Versus Unranked P5
Home Win 3
Home Loss -1.5
Neutral Win 4
Neutral Loss -1.0
Away Win 5
Away Loss -0.5
Versus Ranked Team
Home Win 5
Home Loss 0
Neutral Win 6
Neutral Loss 0
Away Win 7
Away Loss 0
We nees to bring back the poll and just let people vote like the AP poll. 25 points for 1st, 24 for 2nd, etc and tally them up for a top 25.
RE: Working on something and need input - banker - 09-16-2016 01:01 PM
(09-16-2016 12:39 PM)pilot172000 Wrote: Now, does the team have to be ranked when you play them or can they be like a week or two later?
It will based on when you play them.
RE: Working on something and need input - ThreeifbyLightning - 09-16-2016 01:04 PM
Given that any given year half the teams that start the season ranked in the top 25 don't actually finish in the top 25 you end up with a terribly flawed and subjective analysis that favors teams that are lucky enough to have faced a team early when they were ranked and shouldn't have been or face them late in the season when they weren't ranked for most of the earlier part of the year when other teams played them and didn't get the benefit of the points boost. Take the CMU/Ok State game for example. No way Ok State is worthy of a top 25 ranking but CMU is going to get a points boost as if they were a top 25 caliber team. Big flaw.
I won't begrudge you for coming up with a ranking system. Knock yourself out, but don't pretend like it's objective. FWIW, you did ask for input.
RE: Working on something and need input - stinkfist - 09-16-2016 01:07 PM
would only be valid after the season as proposed....
it's why preseason rankings and polls are a fk'n joke until mid-season
sorry....you asked....
RE: Working on something and need input - MU ATO - 09-16-2016 04:20 PM
(09-16-2016 10:57 AM)banker Wrote: I am working on a ranking system (objective) for G5 programs and am developing a scoring system. Here's the rough start, but I would like some input.
Versus FCS
Home Win 1
Home Loss -3
Neutral Win 1.5
Neutral Loss -2
Away Win 2
Away Loss -1
Versus Unranked G5
Home Win 2
Home Loss -2
Neutral Win 2.5
Neutral Loss -1.5
Away Win 3
Away Loss -1
Versus Unranked P5
Home Win 3
Home Loss -1.5
Neutral Win 4
Neutral Loss -1.0
Away Win 5
Away Loss -0.5
Versus Ranked Team
Home Win 5
Home Loss 0
Neutral Win 6
Neutral Loss 0
Away Win 7
Away Loss 0
Couple issues right off the bat.
1. Points for playing 1-AA. They are irrelevant. The only time it's bad is when you lose. So you shouldn't get rewarded for playing or winning vs a 1-AA.
2. You're falling into the Roy Kramer / BCS mode by automatically giving a "POWER" 5 school credit over a NON "power" 5 based on your points. That's nonsense.
3. You rankings as a whole make Roy Kramer & Co. laugh at you, me, and everyone else they have been laughing at since day 1.
RE: Working on something and need input - banker - 09-16-2016 07:08 PM
Well, MU ATO, when G5s start winning more than 15% of their games against P5 schools Kramer will be wrong and I'll adjust the scoring.
RE: Working on something and need input - banker - 09-16-2016 07:19 PM
Here's another one if you don't like the scoring model, outside of Kansas, who just about anyone can beat, take what you consider the bottom 20 G5 schools and the bottom 20 P5 schools, match them up and tell me how many G5 wins you really see. Do the same for the middle 20 and the top 20. Which is harder, beating Charlotte or beating Purdue? Which is harder, beating Boise or beating Alabama? How about wvu or ODU?
I don't understand why some people want to act like there is no difference then the same people get excited when we beat a mediocre, 7-5 Maryland team in a bowl or beat a crap team like Purdue in Huntington. If there is no difference, how does the G5 only muster about 20 wins a year against those teams out of 120 opportunities?
I don't know if people are embarrassed to admit it or what. There's no shame in being behind them given the overwhelming advantages they have.
RE: Working on something and need input - goherd24herdfans - 09-16-2016 08:59 PM
Banker, because typically those G5s are playing body bag games against top 25 competition on the road.
Maryland is a mid tier p5, generally.
Schools like Kansas, Purdue, Wake, etc are always bottom tier p5 and no better than upper mid to top tier g5 schools.
Just feedback man. Point systems like this are bogus. This one in particular is almost an exact rip off of "move the needle". I appreciate the thought and effort into coming up with an objective system, but this is not the way to do it. It has no regard for SOS at all.
RE: Working on something and need input - goherd24herdfans - 09-16-2016 09:00 PM
A rare occasion when i agree with mu ato haha
RE: Working on something and need input - CoachMaclid - 09-16-2016 09:15 PM
Reading this, I think I'd ask what the goal or objective of this rating is? There are 80 different rating systems that you can find in the composite ratings that are cutting things many different ways that are build to be predictive and weighted in various ways. What you have listed is more of an "achievement score" rather than apply any kind of model, so I don't know what question your ranking of points answers?
RE: Working on something and need input - banker - 09-16-2016 09:30 PM
Maclid, you are correct, it's simple keeping track of what has happened, not saying who the best or worst team is. I think it is much more meaningful on a conference level because that eliminates almost everything people think make the results meaningless. When you look at the 44 to 52 non conference games each conference plays, you will get a very similar mix of opponents. So it is essentially an apples to apples comparison of the conferences.
On an individual team basis, the harder your schedule, the more opportunity for points but also the lower likelihood of getting them.
RE: Working on something and need input - banker - 09-16-2016 09:57 PM
(09-16-2016 08:59 PM)goherd24herdfans Wrote: Banker, because typically those G5s are playing body bag games against top 25 competition on the road.
Maryland is a mid tier p5, generally.
Schools like Kansas, Purdue, Wake, etc are always bottom tier p5 and no better than upper mid to top tier g5 schools.
Just feedback man. Point systems like this are bogus. This one in particular is almost an exact rip off of "move the needle". I appreciate the thought and effort into coming up with an objective system, but this is not the way to do it. It has no regard for SOS at all.
Less than 30% of G5/P5 match ups could be considered body bag games.
The only commonality between what I am doing and the move the needle thing is that they both involve college football and the awarding of points. Their system would give you more points for beating Kansas at home than beating Houston on the road (4 more points). In my system, the Houston win would be worth 2 more than the Purdue win. I also penalize you for losing games you should win, or have the opportunity to win. I have also significantly contracted the point differential between G5 and P5 wins. The mere fact that they draw a distinction between a G5 ranked team and a P5 ranked team is ridiculous. A ranked team is a ranked team.
Here's what I believe, after the OOC season wraps up, if you look at the list I post you will be 90% in agreement with the order that is produced. Keep an open mind. It will also be interesting to see how it goes from season to season. Conference point records, individual team records, etc.
RE: Working on something and need input - stinkfist - 09-16-2016 11:51 PM
(09-16-2016 08:59 PM)goherd24herdfans Wrote: Banker, because typically those G5s are playing body bag games against top 25 competition on the road.
Maryland is a mid tier p5, generally.
Schools like Kansas, Purdue, Wake, etc are always bottom tier p5 and no better than upper mid to top tier g5 schools.
Just feedback man. Point systems like this are bogus. This one in particular is almost an exact rip off of "move the needle". I appreciate the thought and effort into coming up with an objective system, but this is not the way to do it. It has no regard for SOS at all.
this is where you and me agree.....there are now 120 whatever the fk teams that play 12 games/yr....
statistically speaking, it's like asking the maid to tune up the engine....
I'll never understand how those try to define in this scope....