don't mind me, I'm just pointing out the flaws and the reality of it.

(04-30-2013 11:02 PM)_sturt_ Wrote:  

(04-30-2013 10:57 PM)randaddyminer Wrote:  there were 3 teams in the mwc that ranked almost equal or better than the aac's #1 last year?

Average of five years... attempting to measure the sustained strength of a program over time.

For instance, how can you rate ODU when they've been a DIAA team? UTSA has only been around how many years now?

I still think that's the one flaw that Sagarin's had in their "5 year average". 5 years is NOT enough to come up with a reasonable average for success/failure, as there's simply not enough data to work with. Not to mention your success/failures can be skewed with just one really bad year

Not to pick on the schools involved, but I'm going to use La Tech (incoming)/USM(steady)/SMU (outgoing) as the guidelines:

La Tech - Past 5 years (3 winning seasons, two losing)
2012 9 3
2011 8 5
2010 5 7
2009 4 8
2008 8 5

USM - One really bad year... and 4 winning seasons.
2012 0 12
2011 12 2
2010 8 5
2009 7 6
2008 7 6

SMU - outgoing (one really bad year, one average, two good)
2012 7 6
2011 8 5
2010 7 7
2009 8 5
2008 1 11

SMU - 31 wins/34 losses, USM - 34 wins/31 losses, yet SMU gets 3 points better than USM?

La Tech has had fewer winning seasons than USM, but are tied for wins 34, USM has 3 more losses (might be bowl losses, but would have to check the schedules to be sure)

Again, there's just not enough data in a 5 year average and sometimes I find their numbers a little iffy, as one year can completely skew your standings. Now if KenPom did a valid football study of some sort, or if we knew how Warren Nolan came up with their NPI's... those I'd be interested in seeing.

I believe Mr. Sagarin would answer that mathematically, once enough teams have played each other and you have those results, there is a network of numbers created that can then be plugged into his calculation, and which then yields a number that establishes a given team's strength... which, in turn, can be used to predict winners and losers with 76% accuracy.

And specific to the question of how FCS programs get measured, FCS has, for as long as I can remember (and that goes back to Marshall's SoCon days) been counted in the Sagarins. Evidently FCS and FBS play each other just enough that the math works.

Certainly, anyone can pot shot how a calculation is constructed, and there will always be... ALWAYS be... the "yeah-but-what-about-this-and-that." Every year brings its own exceptional events that someone might call "flaws," and that typically require some amount of subjectivity to explain them away.

But if given the choice of measuring according to my subjective take, given my biases, and measuring according to sterile numbers that have no bias, any rational person is going to choose the latter.

And one more thing, relevant to the case you've made Saint... I think Mr. Sagarin would point out that the most accurate of the his calculations (there are actually three different ones if you have looked into it) includes margin of victory, which bring into the equation more precise numbers than simple wins and losses.

(04-30-2013 01:16 PM)_sturt_ Wrote:  

(04-29-2013 09:35 AM)baruna falls Wrote:  

(04-29-2013 08:31 AM)_sturt_ Wrote:  In terms of on-the-field production, the five-year Sagrin averages say MAC is well behind CUSA 2.0, but has a slight edge on CUSA 3.0.

thanks, that is what I should have said.

No prob... here's the actual chart, baruna, fwiw...

Wait, who produced this chart? I don't see a link, and when I do a google search I can't find it anywhere. Please provide a link. There are too many discrepancies in this chart for the ranking. Especially with respect to the still-FCS teams and where they would fit in the Go5. On top of this, Sagarin specifically states that head-to-head is the most significant factor in his calculations. Yet, (as I have pointed out), Idaho is ahead of several new teams that have beaten them just last season. If there isn't enough data to make the calculations, then it's quite apparent that new teams will always be ranked low even if they go undefeated according to your chart.

There are just too many discrepancies. Please provide a link where we can read the footnotes and read the rationale.

(05-01-2013 10:33 AM)slow-runner Wrote:  Wait, who produced this chart? I don't see a link, and when I do a google search I can't find it anywhere. Please provide a link. There are too many discrepancies in this chart for the ranking. Especially with respect to the still-FCS teams and where they would fit in the Go5. On top of this, Sagarin specifically states that head-to-head is the most significant factor in his calculations. Yet, (as I have pointed out), Idaho is ahead of several new teams that have beaten them just last season. If there isn't enough data to make the calculations, then it's quite apparent that new teams will always be ranked low even if they go undefeated according to your chart.

There are just too many discrepancies. Please provide a link where we can read the footnotes and read the rationale.

I'm happy to send you the Excel file, unless you'd like to work it up on your own... it took me about an hour on a Sunday afternoon about a month ago. (Not sure how much "rationale" one needs... it's merely averages of rankings over five years. The only footnotes have to-do with a program like UTSA that hasn't had a full five years of football.)

My opinion is that there is some validity to your criticism, but... if I may be so bold... that you're over-stating it somewhat, perhaps because it doesn't reach the conclusions you subjectively think it should reach.

If I were doing it over again, it's occurred to me that it might be improved as a rating of overall strength-of-program to build into the equation some weighting so that the most recent season counts more than does the season before it, and then that season counts more than three years back, and so on through year 5.

And there is no magic, per se, to using five years, but I think that's a long enough period of time to iron out any volatility, yet short enough that you aren't bringing in too much history into the equation.

The fact that App State is at the top of the Sun Belt ranking removes any validity from the data. You can't compare them playing an FCS schedule to Troy and ASU a full FBS schedule.

(05-01-2013 10:51 AM)_sturt_ Wrote:  

(05-01-2013 10:33 AM)slow-runner Wrote:  Wait, who produced this chart? I don't see a link, and when I do a google search I can't find it anywhere. Please provide a link. There are too many discrepancies in this chart for the ranking. Especially with respect to the still-FCS teams and where they would fit in the Go5. On top of this, Sagarin specifically states that head-to-head is the most significant factor in his calculations. Yet, (as I have pointed out), Idaho is ahead of several new teams that have beaten them just last season. If there isn't enough data to make the calculations, then it's quite apparent that new teams will always be ranked low even if they go undefeated according to your chart.

There are just too many discrepancies. Please provide a link where we can read the footnotes and read the rationale.

I'm happy to send you the Excel file, unless you'd like to work it up on your own... it took me about an hour on a Sunday afternoon about a month ago. (Not sure how much "rationale" one needs... it's merely averages of rankings over five years. The only footnotes have to-do with a program like UTSA that hasn't had a full five years of football.)

My opinion is that there is some validity to your criticism, but... if I may be so bold... that you're over-stating it somewhat, perhaps because it doesn't reach the conclusions you subjectively think it should reach.

If I were doing it over again, it's occurred to me that it might be improved as a rating of overall strength-of-program to build into the equation some weighting so that the most recent season counts more than does the season before it, and then that season counts more than three years back, and so on through year 5.

And there is no magic, per se, to using five years, but I think that's a long enough period of time to iron out any volatility, yet short enough that you aren't bringing in too much history into the equation.

This explains a lot. It does not comport to the final 2012 Sagarin ratings:

http://usatoday30.usatoday.com/sports/sagarin/fbc12.htm

I realize you are trying to use a 5 year average. But over the last 5 years I can see that Georgia Southern has been on a tear yet is placed well behind Appalachian State. Add to this the rankings of Idaho vs UTSA/TxST/etc with very recent real-data, and things don't seem to fit. I believe if you are sincere in your analysis, then you are using a time-frame that is too long (I'm referring to a 5 year average).

As for what I desire, I desire accurate sources and accurate information.

If you have a model, great. Is it purely mathematical? If so, great. But is Appalachian State really better than Georgia Southern over the last few years? Is Rice really behind Appalachian State? Is LaTech really above Utah State and SDSU (which both beat LaTech last season)?

Averages are fine but it takes several seasons to move those averages if you use too long of a time-frame. This means those 5 year averages do not accurately capture changes to programs.

You could use a 2 year average or a 3 year average. Five is too long. I would also recommend adding columns indicating where each team fell within the entire rankings instead of adjusting them for Go5. Then do a sort for just 1 year, then a 2 year average, then a 3 year average. Then we have a model based solely on Sagarin and no one will question your data; instead they will question Sagarin's data.

If you did this, then we could see how programs are changing season-to-season.

Thanks for clarifying the chart. I would love to see it based on smaller time-frame averages, as well as just based on last season.

I checked his math for ODU and came up with the same but I agree with Slow Runner and W49r's points. 2012 was ODU's 4th year, our ending sagarin ratings have been 199, 168, 123, 117 so we're obviously trending up and I hope we don't fall as low as our 151 4 year average in the coming years. Which brings us to Wh49r's point of comparing FCS to FBS programs. I beleive there's some overlap b/w the top of the FCS and bottom of the FBS but I don't think there's an FBS team that should be ranked in the 150's. With 125 FBS teams that would put 25 FCS temas in front of them. The #25 ranked FCS team at the end of 2012 was Eastern Illinois, are they really better than NMSU or Tulane? I doubt it.

(05-01-2013 11:09 AM)wh49er Wrote:  The fact that App State is at the top of the Sun Belt ranking removes any validity from the data. You can't compare them playing an FCS schedule to Troy and ASU a full FBS schedule.

To that, I think Sagarin (or any of the other stats-geeks who have constructed calculations employed in rendering the BCS ratings) would say that the numbers don't look at labels... they simply look at results... and again, once you create enough interconnections of margin-of-victory results from all of these teams, you end up with an unbiased estimation of program strength that is substantive enough to actually predict wins and losses.

That's the scientific answer.

Then, my biased answer as a fan is, hell yes, some of the upper crust FCS programs have had stronger programs over a five-year period than some of the lower-echelon FBS schools. At Marshall, we know something about that, having come into the MAC and... much to the surprise of the MAC fans who thought the 1-A and 1-AA labels meant something... not just competed, but dominated.

And Appy? Appy has always been strong. Surely you remember what they did to Michigan. Sure, that was surprising to everyone, but a whole lot less surprising to Marshall fans.

(05-01-2013 11:58 AM)mturn017 Wrote:  I checked his math for ODU and came up with the same but I agree with Slow Runner and W49r's points. 2012 was ODU's 4th year, our ending sagarin ratings have been 199, 168, 123, 117 so we're obviously trending up and I hope we don't fall as low as our 151 4 year average in the coming years. Which brings us to Wh49r's point of comparing FCS to FBS programs. I beleive there's some overlap b/w the top of the FCS and bottom of the FBS but I don't think there's an FBS team that should be ranked in the 150's. With 125 FBS teams that would put 25 FCS temas in front of them. The #25 ranked FCS team at the end of 2012 was Eastern Illinois, are they really better than NMSU or Tulane? I doubt it.

I've already addressed most of this with my previous post... but with regard to the trending thing... that's what made me think that the weighting would be a good addition to the calculation.

(04-30-2013 10:10 PM)randaddyminer Wrote:  

(04-30-2013 10:03 PM)AirRaid Wrote:  

(04-30-2013 09:01 PM)randaddyminer Wrote:  the mwc is A LOT better than the aac, sugar-coat it however you want, the mountain west > murican market conference and it ain't even close.

So delusional. Just because you got left out does not mean you can simply spout non sense.

quit your crying, we're all chasing the Mountain West

MWC wants us, they don't want you.

(05-01-2013 04:03 PM)AirRaid Wrote:  

(04-30-2013 10:10 PM)randaddyminer Wrote:  

(04-30-2013 10:03 PM)AirRaid Wrote:  

(04-30-2013 09:01 PM)randaddyminer Wrote:  the mwc is A LOT better than the aac, sugar-coat it however you want, the mountain west > murican market conference and it ain't even close.

So delusional. Just because you got left out does not mean you can simply spout non sense.

quit your crying, we're all chasing the Mountain West

MWC wants us, they don't want you.

nobody wants a fake ass school, whose big donors, gotta buy tickets to have a sell out of a 30k stadium... Move along you horn/aggie

(05-01-2013 07:28 PM)randaddyminer Wrote:  

(05-01-2013 04:03 PM)AirRaid Wrote:  

(04-30-2013 10:10 PM)randaddyminer Wrote:  

(04-30-2013 10:03 PM)AirRaid Wrote:  

(04-30-2013 09:01 PM)randaddyminer Wrote:  the mwc is A LOT better than the aac, sugar-coat it however you want, the mountain west > murican market conference and it ain't even close.

So delusional. Just because you got left out does not mean you can simply spout non sense.

quit your crying, we're all chasing the Mountain West

MWC wants us, they don't want you.

nobody wants a fake ass school, whose big donors, gotta buy tickets to have a sell out of a 30k stadium... Move along you horn/aggie

LOL, the butthurt is strong. I won't rub salt on your wounds.