Quote for Zoomie..

[Don]

re: Sagarin Elo

"Often people who are not familiar with the nature and limitations of statistical methods tend to expect too much of the rating system. Ratings provide merely a comparison of performances, no more and no less. The measurement of the performance of an individual is always made relative to the performance of his competitors and both the performance of the player and of his opponents are subject to much the same random fluctuations. The measurement of the rating of an individual might well be compared with the measurement of the position of a cork bobbing up and down on the surface of agitated water with a yard stick tied to a rope and which is swaying in the wind."
Arpad Elo in Chess Life, 1962

This from the "inventor" of the "ELO" system used to rank Chess players and (through Sagarin) now football teams.

[Bucs90]

Wow. From ranking chess players to college football.

"Comparison of performance". That is a very strong statement. Think about it. A "performance", meaning a full season worth of performance. Meaning CRUNCH TIME performance as well as performing while being heavily favored.

What Zoom doesn't get is simply wins and losses are but one peice of a comparison of performance. The ability to win in crunch time trumps the ability to simply win. Missouri, Kansas and OU all proved they can win. Missouri and more so OU proved they could win in crunch time. Kansas could not.

[CJHawkeyes]

Wow. From ranking chess players to college football.

"Comparison of performance". That is a very strong statement. Think about it. A "performance", meaning a full season worth of performance. Meaning CRUNCH TIME performance as well as performing while being heavily favored.

What Zoom doesn't get is simply wins and losses are but one peice of a comparison of performance. The ability to win in crunch time trumps the ability to simply win. Missouri, Kansas and OU all proved they can win. Missouri and more so OU proved they could win in crunch time. Kansas could not.


Comparison of performance does not imply crunch time or performing when heavily favored. Performance comparison can simply mean comparing a 12-1 record to an 11-3 record. The quote says everything Zoombag and I have ever argued. The people that Sagarin thinks expect too much of a statistical ranking system are those that expect it to tell them who is the better team. Again, for some reason I haven't figured out, you seem to think what teams do in an eight game performance automatically trumps what the same teams do in a 12 game performance despite the performances being judged by different rules just because the first performance passes out rings to the winners.

[Bucs90]

I just place a far bigger emphasis on crunch time. If a team chokes in the crunch, it IMO trumps anything they did before. Colts went 13-3. Choked. Their season ruined. Kansas went 10-0. Choked. Season lost.

Thats just my opinion. My view on college football is that ALL of the OOC and regular season is nothing more than a means to qualify for a chance to win the conference, or to win it outright if there isn't a CCG. And then, from that point, we see who qualifies for the BCS game, although I hate the BCS system.

Some guys don't agree with me, thats fine. From a pure competitive standpoint, you gotta win the conference. That trumps all else that occurred prior. Just my opinion, and that probably will never change. Sports are meant to come down to those glorious occassions when all the marbles are on the table, all the work and sweat was just to get to that point, when all the fans are on the edge of their seats if sitting at all. Missouri and Kansas both earned their way into that unofficial North Division Championship game. What occurred before that day was nothing more than the quest to get there.

[CJHawkeyes]

It is fine if you hold that position. However, your position does not mean someone else cannot place more value on the entire season based on one set of rules just because another competition favored different teams based on a portion of the season under another set of rules.

[Bucs90]

True, but I can also point out why I feel that is one of the worst ways to rank teams. The basis for debate.

[CJHawkeyes]

True, but I can also point out why I feel that is one of the worst ways to rank teams. The basis for debate.


Of course you can, but your argument applies to more possible rules than just going by records. The bottom line is you are placing more value on the results of an eight game competition under one set of rules than on a 12 game competition under another set of rules. Besides that, Zoombag's argument does not suggest that records trump playoff results. The Colts were seeded higher than the Chargers due to records. In playoffs, records cease to matter. However, in college football, no Big 12 team competed in a playoff. Therefore, ranking teams by records is a reasonable position to take.

[ZOOMBAG]

Yea, I fail to see any valid point as it concerns using rating systems to rate teams when there are no stated rules of competition at a national level to determine a champion.

Rating a 120 teams on 12 or 13 games where many teams have identical or near identical records and the rules do not prevent inequities in scheduling, we have to statistically insignificant factors to split hairs.

But if one has to "rate" 120 team in order to pick a #1 and #2 because no system exists to equitably determine that on the field one has to resort to "rating" teams. And if so, it has to be done objectively and in accordance with valid statistical methodology. The result is that a Sagarin can confidently predict that his #1 team playing his #2 team on a neutral venue 99 times that #1 would win at least 50 times by at least one point, even if they might lose 49 by 10, 20 ,or more....but if you happened to observe only one the 49 losing events the typical idiot would claim the system was invalidated by the one event....

[ZOOMBAG]

A "performance" is nothing more than a collection of statistics. Points scored, points given up, yards gained, yards given up, etc.... They can be measured and collected and analyzed and given enough events a valid statistical model can be derived.

BTW, Kansas won the Orange Bowl....in "crunch time". So add "crunch time" stats to "goals met" for the Bucs90 rating system....

[ZOOMBAG]

All you have is a hypothesis. In order to get anyone to seriously buy it as a statistical model, you have to mathematically PROVE the weighting factor you give it (a number you have yet to divulge...).

[GatorGrad]

Didn't LSU finish #1 in Sagarin's final rankings?

[Blue Hen]

They did, indeed. LSU finished 0.61 point ahead of Kansas.

[GatorGrad]

So why is Zoom using Sagarin to support his claim that Kansas should be the 2007 Mythical National Champions?

[Blue Hen]

don't know. Haven't been following this debate .