Sagarin’s preseason rankings

He’s got Georgia ninth, LSU second, Michigan eleventh and Arkansas twenty sixth, for starters.  And he’s showing Southern Cal as the dominant team in the game this season, with a huge spread between it and LSU.

Since his numbers are power rankings to which you would add 2.87 points for home field advantage, he’s calling for Georgia to lose three games in ’08, to LSU, Florida and Auburn.

He doesn’t have strength of schedule numbers posted, but he shows Georgia playing #116, #85, #62, #46, #43, #32, #31, #21, #13, #10, #5 and #2, which seems pretty respectable at this juncture.  Take it for what it’s worth.

About these ads

14 Comments

Filed under College Football, Georgia Football, Stats Geek!

14 responses to “Sagarin’s preseason rankings

  1. Chuck

    Michigan at 11? Where did he have Notre Dame last year preseason?

  2. Chuck

    In fact, considering that that’s a power ranking and not a prediction of finish, I find a lot that’s puzzling in there.

    We’re a worse team than Virginia Tech? West Virginia? We’d be touchdown underdogs at LSU? Underdogs at Michigan?

    I’m really not seeing it.

  3. It seems like how teams finished last year has what Donald Trump would call a yooooge influence on his rankings now. That’s the only explanation for Michigan (final ’07 ranking: #18/#19) and Arkansas (#31/#40) being ranked as highly as they are.

    These computer rankings tend to fluctuate wildly in the early part of the season as actual games start taking place, so Sagarin’s preseason list should be taken with a Costco economy-sized grain of salt.

  4. Then what’s the explanation as to why Kansas, his #2 team at the end of last season, is a relatively lowly #29 in the new ratings?

    I agree with you on the grain of salt, by the way. I just can’t figure out where he came up with these numbers.

  5. NM

    Probably the same place Playboy came up with theirs…

  6. Hobnail_Boot

    I also like getting hammered and throwing darts at a wall whilst blindfolded.

    Michigan? Why the hell not? Let’s do shots!

  7. Ally

    *Threadjack*

    Senator, did you see this today? It was the Hawt post:

    http://djrebel.wordpress.com/2008/08/20/fyi-yes-it-really-happened/

    Just thought you might get a kick out of that.

  8. I saw it – Orson over at EDSBS linked to it. Why am I not surprised?

  9. I can’t find a better link to back it up than this, but it’s my recollection that the initial Sagarin ratings use the final rankings from the past several seasons (as many as the last 4 or 5).

    As the season goes on, there are enough points of comparison from the current year that his preseason ratings are gradually phased out before the first BCS rankings.

    If so, that would explain 1) why SoCal is so high and dominant (not many other programs have been as solid over the past five years and the ones who have been – LSU, Ohio State, and Oklahoma – are right up top too), 2) why Kansas is so low (one year disappears quickly when averaged into their recent past) and 3) why Georgia is so meh (they’ve finished anywhere from #2 to barely in the top 25 lately).

  10. Sam

    Sagarin has traditionally been one of the “computer polls” that seem to have the most variation from the people polls. I always thought he had a West Coast bias as the PAC 10 teams had very strong ratings even when they lost outside the conference. This was because they played such highly rated teams within the conference due to the aforementioned bias at the beginning. I am not a conspirarcy guy, just think Sagarin has a huge bias and the starting point never gets overcome. I always felt his poll was pretty weird. A couple of years ago he had some teams out West rated so high it was comical. I would guess a highly inflated USC start point would distort the relative values of the PAC 10 again.

    Theoretically, shouldn’t all teams start with the same power rating, say 100, and let the interactions once, twice, thrice, etc. removed play out? Only a computer can do that as it gets very tangled after a few weeks, but I thought this was the purpose/strength of the computer polls. It is the assignment of relative values going into the season that does Sagarin in.

  11. This is my problem with having computer polls play such a big part in the BCS standings. Earlier the Senator said, “I just can’t figure out where he came up with these numbers” — well, nobody can (except Sagarin).

    The assumption is that computer rankings are pure, cold, hard numbers and, as such, are totally objective, but that’s not necessarily the case. I mean, who knows what kinds of weightings Sagarin is giving to the various kinds of data he’s using in his rankings? Not me, and even if I did, I’m sure I wouldn’t be able to understand them unless I had an advanced calculus degree.

    The pro-computer folks point to rankings like Sagarin’s as cold, unbiased numbers that balance out the biases and personal prejudices of the human voters, but in actual practice there’s far less clarity or transparency with the computers.

  12. Sam

    The fact that ALL computer rankings aren’t virtually identical would indicate it is not truly a scientific method. I am not saying it couldn’t be if the starting point for all teams were the same and the interaction played out, but for now they look to be as subjective as the other polls with someone’s agenda distorting the rankings.

  13. TexasRangers

    Georgia is ranked #9 WHAT?! Must be a computer glitch. That’s insane! Satan Sagarin can burn in computer hell.

  14. Francis L. Kulchar

    Can I get the Sagarin OPENING RATINGS for each NFL team for 2008, prior to Sept. 4??

    Thanks,

    FLKulchar