And the SEC currently has 11 of 14 teams bowl eligible.

If Tennessee beats Vandy and Kentucky beats Louisville then the SEC will have 13 of 14 teams bowl eligible.

Based on this and krpinds points...

So not necessarily running up the score then. Because the margin might not tell the tail of the game or be particularly indicative of simply trying to run up the score.

At what point then, do you think it becomes obvious that the margin shows that the losing team just isnt very good?

What I'm getting at is. Suppose everyone can put a margin of 50 points on team SucksU, if HighfligherU puts a margin of 100 on them, but only puts an average margin of 10 points on everyone else.  An aggregate of the margins gives WAY too many points  from SucksU, because well, they suck. If you take the aggregate of a team that didnt get to play SucksU, they are at  huge disadvantage in the "total margin" comparison even if their avg margin over everyone else was 15pts a game. Putting up over a certain amount of points, I dont think necessarily says anything about the win. Losing by a margin of 50 and losing by a margin of 70 both say you suck and at that point I dont think it says anymore about how much you suck.   EI diminishing returns.


IMHO anything over 40 points of margin indicates just a huge mismatch and isnt necessarily telling you anything more.


Or do you simply say, who gives a shit about your margin against teams with losing records? Then, you dont even have to get into what that "too much" margin should be.

Yeah, but that'll never happen cause the SEC sucks.  I know it's true cause I read it right here.  

Not unexpected, in my book. Bret Bielema is an ass, personally. But the guy can coach, and the players respond to him. His first couple years as the head coach in Madison were mediocre - not nearly as bad as last year in Ark, but look at the talent Bret inherited when he took over the Badgers. After he figured out what he was doing, the teams got good. Really good. Not national top tier good, but consistent challengers for conference titles. I don't care for the man, I don't like the way he walked out on Wisconsin, but I think he will have some success. Arkansas has the advantage of being able to pay higher than Wisconsin did for assistants, which will help him out (one of the things he was unhappy about in Madison). I think Arkansas will be a perennial bowl team going forward. But I won't root for them.

I think it's shittier to let up on a team than to run up the score.

Find that quote or kindly admit you're a making shit up....kthx.

How many are your team?  Or do you just root for a conference ?

Really?  I'll stop posting in this thread for a week if you aren't just lying.

Conference homers are something else.


Goooooo Teams!

My angle is more along the lines of, at what point does scoring another TD not really tell you anything more about the difference between the skill levels of the teams?

Real life example. Marshal is currently ranked #1 in the nation in average score margin. However they haven't played a single ranked opponent. Havent even played a single Power-5 opponent.  Hence, margin means dick.

For those looking for a good laugh brought to you by some delusional Ohio State homers, enjoy:

http://www.ar15.com/forums/t_1_5/1563091__ARCHIVED_THREAD____Roll_Tears_Roll___.html&page=2

A small excerpt of the hilarity from a post by Infantry26:


Link

Yes. I think in some cases scoring means nothing. However, I think there are some things that are directly comparable. Scoring however, along with a whole host of other metrics, can be used for rank weighting. Scoring itself, can not be used for prediction. For prediction, of the various metrics out there; I like offensive and defensive efficiency. However, even that, isnt weighted for cream puffs or cross conferences. Offensive/defensive effeciency doesnt mean much, if your stats only come from creampuffs. To put this in abit of context. One of the weighting metrics that I've found for place ranking (not prediction), that is used in combination with other ranking metrics, is a bined quality of win/quality of loss. A win vs a team with a high w/l ratio is worth more than a win compared to a team with a low w/l ratio. Likewise a loss vs a team with a low w/l ratio is more "damaging" compared to a loss vs a team with a high w/l ratio.

However, this is different than the basic crosscorrelation ive been working on. Anyone can apply weighting factors and come up with a rank. I was just curious on folks opinions on margins.

The bigger picture on this.

All of the ranking schemes out there use a variety of metrics for placement, each metric in itself having a certain weight. eg - some systems place more emphasis on # of wins, others types of wins, others margins. However, every single one of them. ALL of the computer models, fucking everything, uses polled data in some or alot of their metrics. So these "computer" rankings, still rely on human bias. This, is mainly for cross conference comparison. I admit, it does have some degree of merit. Without it, you get convergence in local maximums (somewhat talked about in a different post of mine). As in if you dont get alot of cross conference play, you can easily be #1 in wins, and strength of schedule, for your conference.  But how do you really compare a undefeated vs an undefeated if they have no common opponents? What about an undefeated vs a 1 or two loss team? Until this point its pure human opinions. Fucking conjecture.

Now, I'm not trying to come up with new weighting factors. Those have been come up with ages ago. SOS, powerfactor, margin, etc. My last couple of posts have simple been asking what yall though a reasonable cut off point for margins would be. As in a very small subfactor weighting, leading to an overall ranking. It's also a purely subjective ranking as well. You can weight factors however you want. What I'm working on, and is the most pivotal part, is that conference to conference comparison. Pattern recognition, in a relative math sense, is still in it's infancy. The oldest of techniques wasnt even developed until the mid 80s, and realistically hasnt been used until very recently. Most of the techniques relies on huge matrix operations. Pulling an eigenvalue or eigen vector of a matrix larger than a 3x3, by hand? FUCK THAT! Hell, when I first started doing this stuff in ~05-06 the common computing programs would choke on pulling eigenvalues from large arrays. Performing those operations on a 128x128 matrix (FBS teams), simply wasnt going to happen. Essentially I am not surprised in the slightest I havent seen a single pattern/group cross correlation done on FBS teams.

Now, if you can find the correlation coefficient between teams or conferences. Bam! You instantly got rid of human opinion about conference/conference.  You know exactly how, Clemson compares to Alabama, or how Arizona State compares to Michigan.

That's not ranking mind you, the correlation coefficient simply tells you just how those groups are like each other. As in, if they were in one big group together, how appropriate data compares between them. But if you know the degree to which they are alike, you can now apply weights to in conference and out of conference games. Convergence to local maxima based on SOS, Power, quality wins/loss, suddenly becomes a thing of the past. You have a unified cross correlation weighting between conferences. From there, you can apply your own ranking criteria and weight it however you want.

To reemphasize, I dont claim any of this to be predictive. In college ball, anything can happen. Being the Gator homer that I am, I never would have guessed the result of the UF/UGA game. I thought they might keep it close, that's it.

However I do think I can find a better, non-biased ranking, for cross conference comparison. Find the best, proven performance, to get the actual best teams into the finals.

Well it's that time of year
ETA: Deleted. I can't do it, it isn't even fun anymore to tease michigan
 

You misunderstand. Or, I didnt explain it well enough. It's not an attempt to quantify, performance. As you yourself say, that is a very hard thing to do. The weighting metrics for overall ranking, I'll leave to each person to determine on their own.

I'm quantifying relationships between conferences, based on who they played. As in based on who plays who, how they can be compared. The weighting metrics applied after that, that's all personal choice.

As it is, it's easy to compare SOS in an individual conference. Multiple conferences, play teams from other conferences. If a certain amount of teams from conference A, play way more teams from conference B than they do conference C; then you can say conference A and B are much more tightly coupled than conference A and C. Do this for all of the conferences. Based on teams played, if you are likely to identify a team as belonging to conference B, when they in fact belong to conference A. It means those two conferences are very tightly coupled and could most likely be grouped as a single conference. This is true of the east/west, north/south divisions in the various conferences. ACC atlantic and ACC coastal are tightly coupled, so thus should be considered as a single conference. Already I can tell that there are multiple divisions that could be grouped with a different conference at around the same percent as they would be their own conference.  

This allows you to directly apply a weighting between the conferences based on likeness.

Again, this is not the ranking criteria.

It does however allow you to apply a weight to the ranking criteria that you chose, when comparing conferences, based on likeness.




An example not back up by numbers, for illustrative purposes.

If ACC atlantic played as many games against Big 10 east as they do against ACC coastal; you have an direct comparison between ACC at and Big10 east. As in you could group them together as the BigAtlantic. Because the atlantic is also tightly coulpled with the coastal, and the Big10 east is tightly coupled with the west, you could form the supergroup, "BigACC." Teams within this group can be directly compared. It's now very simple to obtain metrics based on overall correlation. If 90% of the time a big 10 team would be confused with an ACC team, based on who they are playing, then a metric comparing the two could simply have a 0.9 weight factor.

If you do this with all the teams, in all the conferences, you'll find various sameness ranks between them all. This is what you can use to directly weight, all your other weights. A scalar for all your other weights. Each division would have a unique scalar for all the other divisions, when recursively applying ranking weights.

A lot of that is predictive. I'm not even getting into that mess.

What I'm actually doing is looking at conference cross correlations. Simply WHO plays WHO.  I dont care who wins, what the score is, how a team does, how many turn overs, anything like that. It simply doesnt mater for what I'm trying to determine. I probably confused the issue by asking about margins and talking about rank weighting in the same post when talking about correlations. Margins has to do with the rankings applied, AFTER you know how teams between conferences are correlated.  

The WHO plays WHO info, lets you apply the very first weighted metric. That metric lets you know by how much the results of one team compare to the results of another team. That metric currently does not exist, outside of a team's own conference. That's what I'm ultimately quantifying. Not performance in the game, or momentum, or if one team had a hard time beating another.  I'm looking at how they are related based on WHO they played.


An easier example to explain. I have three groups; trucks, cars, and boats. To me it doesnt matter how far each of them travel, or how long they can go before filling up. However if mystery vehicle X rolls into the scene, and it has a cab, flatbead, and big dirt tires; I'm very likely to classify it as a truck. The features I'm looking at arent big tires or flatbeads, they are WHO they are playing.

Won't happen but my Hogs would whip FSU ass

Let me know when you shut out two top ten ten teams in a row

Man everytime I watch FSU, I cannot believe the shit they pull and get away with (most of the time)

I'm glad #7 actually got ejected

Keep barely beating crappy teams, and playing like a bunch of thugs..... yeah they deserve to drop.

I turned off the game before the end of the 3rd.  Biggest Wallace shit show I can remember.  Just awful.  I am so damn glad he will be gone next year.  This is normal for Ole Miss, we get our hopes up only to watch the wheels fall off again and again.  I think we have a good thing going for the coming years but damn I get sick of watching talent and effort shit the bed over and over.  Maybe our next QB can be one person for a whole season.  Damn. Fucking hard to even look forward to the egg bowl after this fucking bullshit.

Who did that?

A few questions:

Was it a "live" ball or "dead" ball?

Was the play clock running?

Was the official standing where he is supposed to stand when stopping play? (Hint...its a very specific place from the rule book....and he wasn't there).

Can an official be responsible for "delay of game"?

Who put their hands on who first?

Did you watch the whole game?....did you see a swath of calls that even had the announcers asking "why didn't he call that?"

The official was in the wrong in my opinion.

First off touching a member of the officiating staff is grounds for being kicked out of the game. No matter what your excuse it.

Second, FSU made a substitution, he is supposed to give the opposing team time to make one too, maybe the linemen were assholes and didn't let him through, so he did his job and stopped FSU until Boston College was given adequate time.

If the play clock was running, FSU made a substitution, it is their fault for the time it takes off the clock.... and there wasn't even a delay of game....

Again, shoving a ref several times? Winston is lucky he didn't get tossed out of the game.

Would it soothe the butthurt if you knew the official basically said he was in the wrong?

3:30pm kick off for next weekend, hoping to see some of this.

If UGA is any indication they will show up with their A-game this week.

We need to play mistake free ball.

Skip the fucking formalities! What more do we need to see that Hoke needs to be fired????? The recruiting class is already fucked up, so why not can this fat fucker and already start the search?

This is tough, I think you have to take more into consideration, unfortunately.. a team like Baylor, TCU, Auburn, etcisn't "put away" if you have a 4 TD lead but a team that relies heavily on defense and ball control probably is.

The best way to tell is when the starters start getting pulled, but even then there are many other factors that go into this (quality level of the backups, minor injuries to starters, etc).  In 2012 OSU had to put the starters back in against Indiana.  

Patience, it'll happen. I'd like to see if the Harbaugh rumors turn out to be true.

That would be great for the program.

Courtesy.  We're replacing Bo this year.  You can have next.  

 

Find that quote or kindly admit you're a making shit up....kthx.

The interesting part is that you actually have to take more time and effort to trim the quote that I was responding to out of what I said?  Why bother unless you were trying to take it out of context


You're obviously having trouble deciphering the English language.
LinkedM4 said "FSU would beat OSU by a lot"
I said "They'd do better than the other teams FSU has played"

What part of that even remotely satisfies your accusations in red, killer?

I posted so much last year if I was saying these things or running around saying "SEC sucks, OSU uber alles" or whatever you think was said, it should be all over.

The evidence that you said it is right there and you still deny it.

I guess you think if you start name calling it will take the attention away from your moronic posts that expose your college football ignorance.  Sorry, but it's not working.

http://www.ar15.com/forums/t_1_5/1563091__ARCHIVED_THREAD____Roll_Tears_Roll___.html&page=2

I actually highlighted the part in blue where you started the name calling.  There isn't even a post from me on that page, liar.

You make posts acting as if it was already a done deal that Ohio State was going to play in the BCS Championship game when they had yet to play Michigan State and you believe calling that behavior delusional is "name calling."

Point out where I stated you made a post on that specific page.  You can't because I didn't.  I simply posted the link to that page in the thread because that's where the Ohio State homers started their ridiculousness.  I already posted a direct link to at least one of your posts in that thread.

But since that basic concept confuses you, here:

http://www.ar15.com/forums/t_1_5/1563091__ARCHIVED_THREAD____Roll_Tears_Roll___.html&page=4

I'm still waiting for you to post this quote you keep talking about.  The unedited version please.  Of course, that would just show that you're lying about what was said and showing your ass yet again, killer.

Current lines:

Auburn @ Alabama -7

Mississippi State @ Ole Miss PK

Florida @ FSU -10

LSU -2 1/2 @ Texas A&M

Georgia Tech @ Georgia -14

I realize it's hard for someone like you, but try and keep up. It's already been posted:

http://www.ar15.com/forums/t_1_5/1628999_THE_OFFICIAL_2014_HOUSTON_HUSKER_MEMORIAL_COLLEGE_FOOTBALL_THREAD.html&page=172#i50530280

Despite what you might believe, denying it's existence doesn't make it any less real (especially when the proof is right there for everyone to see).

You said it.  Man up and own up to it.  Or continue making yourself look worse and worse by denying it.  I'm good either way.

Keep him around for one more year; we have to play there next season.  



 

I made it easier for you to understand, liar.

So at this point you're either denying you posted something that I provided a direct link to proving you posted it or you are claiming someone else edited YOUR post in an archived thread to make it look like you posted it.  Either way, you couldn't make yourself look much worse if you tried.

Here it is again because apprently you missed it the previous times it was posted:
LINK

Meh.  Seems to fit the narrative of the vast majority of posts from the last 30 or so pages.  

Fine by me.

I can see how it's getting fairly old as at this point it's pretty much just him repeatedly lying about what he posted despite the fact that proof he posted it was provided multiple times.

Sorry UGA fans, still no line on the Arkansas/Missouri game.