Pollspeak welcomes our computer ranking overlords

Topspin

Verified User
Pollspeak welcomes our computer ranking overlords
Views4638Comments5Posted by: Andy Staples02:15 PM ET 10.06 Share

LSU, darling of the microchip set.::APBy Hugh Falk, Pollspeak.com

It's October, the month where strange and horrific creatures visit our homes and scare the hell out of us. Of course, I'm talking about the computer rankings. The first Anderson & Hester Rankings of the season were just released, and the lone BCS component left to appear is the Wolfe Ratings, which come out on October 18th during a full moon (Yes, that was a lame wereWolfe pun).

More importantly, October is when the computers typically start to make sense. To help prepare you for the BCS horror, let me explain two of the "trickier" rankings: Sagarin and Massey. You can find the Sagarin Ratings at the USA Today site where teams are listed by their standard "Rating." However, this rating isn't used by the BCS, which doesn't allow for margin of victory. Instead, the BCS uses Sagarin's ELO_CHESS ratings, which are listed on the same page in the second column from the right. Unfortunately there is no way to sort by that column. So the best way to see the Sagarin teams sorted in their BCS order is with Pollstalker. The Massey Ratings have a similar "BCS" column (also second from the right). The Massey site lets you sort, which makes it easier to read; however, you must also use the "FBS" filter since Massey doesn't send any FCS (1AA) schools to the BCS. Of course, you could just use Pollstalker to look at the Massey Ratings as well.

The computers definitely see things differently at this point in the season with all five computers tapping LSU as the No. 1 team. Meanwhile all three human polls still rank Florida No. 1. This isn't a surprise, but it does show how differently humans and computers handle their week-to-week rankings. LSU didn't even receive one first place vote in any of the polls (AP, Coaches', or Harris Interactive).

Florida may end up being national champions (relatively good odds according to Vegas). However, everybody who is voting Florida No. 1 is making many assumptions. For example, the Gators will play as well as they did last year. Opponents won't be better at defending the spread. Tim Tebow or other key players won't be injured. Other teams haven't improved more than Florida. Other teams won't get lucky, etc. Computers can't make these assumptions. Instead, they'll have to wait until Florida plays LSU next week. After that, I strongly suspect the computers and humans will be on the same page for the first time this season.

Speaking of those scary humans, shouldn't South Carolina be ranked over Mississippi? The Gamecocks have a better record and beat 'Ole Miss head-to-head. All of the computers rank the Gamecocks higher, but none of the human polls agree. More than half of the AP voters also have it the wrong way around. In fact Mississippi has one of the largest ranges of rankings in the BCS components. The Rebels are ranked No. 16 by the Coaches, but No. 71 by Massey (BCS, FBS). Mississippi gets its chance to make a statement next week when Alabama comes to town. If 'Ole Miss wins, we'll have to wait one more week for South Carolina to play Alabama and sort this mess out.

Meanwhile the Coaches' Poll still has Penn State ranked over Iowa. Iowa struggled with Arkansas State (and won), but they still have a better record and beat Penn State. I'll make the coaches a deal: let us see your ballots, and I promise not to refer to you by name. It can be the way penalties are called, without naming the player who committed the foul. Oh wait, I just remembered….they changed that rule because coaches thought it would be good for everyone to know what is going on even if it did cast a bad light on players. That sounds familiar.

Oklahoma becomes the first two-loss team to be ranked in the polls. There is no shame in that; their two losses were close calls against ranked teams. At least all of the human polls have BYU and Miami over the Sooners this week. That's still not the case with individual AP voters, though. John Clay leads the way with Oklahoma ranked No. 11 (highest) and BYU unranked.

Five AP voters have Ohio State over USC. I'm not sure why, when they lost the head-to-head at home. It must be Ohio State's quality win over…

I mentioned how the computers typically start to make sense in October. One example this week is Doug Lesmerises' ballot. I'm not saying Doug is a computer, but his method of heavily weighing on-the-field results over reputation is computer-like. His ballot this week is his first that isn't the most extreme. Again, we're reaching a point where more and more voters are abandoning reputation over actual results, and Lesmerises is now falling more in line with other voters.

The new extreme voters are Jon Wilner and Scott Wolfe. The California duo was commonly mentioned in past seasons, but not so much this year. Wilner has Miami and Virginia Tech highest, Florida lowest at No. 7 and is one of the few to leave Penn State unranked. Wolfe was the only voter to rank Utah over Oregon, a team who beat them. Wolfe also left Penn State unranked. While I agree that the Hawkeyes should be ranked higher than the Nittany Lions, Penn State is still a quality team. Then again, they are probably still looking for a quality win, and I can't argue with that.

October will be a scary month with both the BCS and Halloween looming. I'm a big fan of both and can't wait to see how the month plays out. Don't worry; Pollspeak will be here to protect you. Just don't believe what BCS Commissioner Jason Voorhees always says, "Relax. It'll all work out in the end."
 
There are always biases. The programmers cannot avoid it unless we every team playing every other team.

I would imagine having 2 games a week and yearround football will never happen.
 
One of the problems with the computer polls is that the computer polls do not use the margin of victory at all.

So barely beating MS State in a last minute goalline stand, and blowing them out so that the game was over in the 3rd qtr, gets you the same results to the computers.
 
I'm not going to argue it, a buddy sent it so I posted it to get a rise out of you guys.

Vegas is probably the best poll, they have the Gators 10pts fav in LSU.
 
There are always biases. The programmers cannot avoid it unless we every team playing every other team.

I would imagine having 2 games a week and yearround football will never happen.
Isn't an opinion stil an opinion even if a computer makes it?

BCS rankings and the polls be damned. You can't have a national championship by Proxy.
 
One of the problems with the computer polls is that the computer polls do not use the margin of victory at all.

So barely beating MS State in a last minute goalline stand, and blowing them out so that the game was over in the 3rd qtr, gets you the same results to the computers.
Yea but Margin of Victory has been debunked. Over the course of a season it's a meaningless statistic. If Texas beats Florida 62 to 7, that means something. If Texas beats Northern Alabama 62 to 7 it means nothing.

Another example. The final score of the Georgia vs LSU game gives no indication of how badly Georgia beat LSU.

So margin of victory is pretty much a meaningless statistic.
 
Yea but Margin of Victory has been debunked. Over the course of a season it's a meaningless statistic. If Texas beats Florida 62 to 7, that means something. If Texas beats Northern Alabama 62 to 7 it means nothing.

Another example. The final score of the Georgia vs LSU game gives no indication of how badly Georgia beat LSU.

So margin of victory is pretty much a meaningless statistic.

Dude are you on crack, LSU was penalized for pointing to the sky.
Does the beatdown to LSU in the National Championship HURT THAT BAD???
 
Dude are you on crack, LSU was penalized for pointing to the sky.
Does the beatdown to LSU in the National Championship HURT THAT BAD???
Oh yea. Give them a penalty after they hand them the game! What's next? A trip to Disneyland for chop blocking? LOL
 
Dude your seriously retarded
LSU won, not the beat down they issued Ohio Slow by a long shot. But a quality win by the 5th ranked team in the Nation. GEAUX TIGERS
 
Yea but Margin of Victory has been debunked. Over the course of a season it's a meaningless statistic. If Texas beats Florida 62 to 7, that means something. If Texas beats Northern Alabama 62 to 7 it means nothing.

Another example. The final score of the Georgia vs LSU game gives no indication of how badly Georgia beat LSU.

So margin of victory is pretty much a meaningless statistic.

Its not the end all stat, but its not meaningless either.
 
Quality wins should matter more.
Saban is not the kind of coach to run up the score.
The system forces you too, Myer on the other hand will call time outs 30pts up with a minute left to try for another TD vs your mom and aunt.
 
Quality wins should matter more.
Saban is not the kind of coach to run up the score.
The system forces you too, Myer on the other hand will call time outs 30pts up with a minute left to try for another TD vs your mom and aunt.

But if two teams play the same team, and one barely survives and one beats them easily, it should count for something.
 
But if two teams play the same team, and one barely survives and one beats them easily, it should count for something.
No it shouldn't. What if it's 30 below zero in a blizzard on the score that's a close game? What if the team that got whooped had it's top 6 players injured during the game? Shit happens. A win is a win and margin of victory means very little as a statistic. Topers right. It just creates an environment where Florida is going to make sure it wins 84 to 3 against Sister Mary's School for Quadraplegics, which means nothing where as beating the top 3 teams in the SEC does!
 
No it shouldn't. What if it's 30 below zero in a blizzard on the score that's a close game? What if the team that got whooped had it's top 6 players injured during the game? Shit happens. A win is a win and margin of victory means very little as a statistic. Topers right. It just creates an environment where Florida is going to make sure it wins 84 to 3 against Sister Mary's School for Quadraplegics, which means nothing where as beating the top 3 teams in the SEC does!

The difficulty of schedule is taken into account. How do you figure that? Beating Bama in 2000 and beating Bama in 2009 are vastly different things. The margin of victory is going to say something about the quality of the teams involved. If Florida lays 50 points on GA, does that not speak to the ability of the Gators? Running up the score is not a sign of sportsmanship, but it is a sign of what the team is capable of doing.

As for having the players injured or having crappy weather, those things effect the outcome of games and no one cares. If Bama loses because 6 of our starters are injured, we don't get an asterisk by the game, we have a loss. If one team is pass oriented and one is run oriented, the run oriented team will win in a blizzard. That does not mean the pass oriented team gets any points for losing in a blizzard.
 
This whole debate is frustrating. If we had a rational play off systems like every other sport we wouldn't even be having this discussion. You win your conference, you go to the playoffs. Come in second, well then maybe polls can be used for filling the rest of the playoff field and seeding it.

If we had a 16 team playoff all 12 Div IA conference champs get an automatic bid. The remaining four teams go to the highest rated teams that didn't win their conference. Screw Noter Dame, they can join a conference. Then use the polls to seed the teams with the top seeded teams getting home field advantage.

Imagine the fun of watching Florida play at Michigan in January for the Championship game? :)

You could play the first 3 rounds the last 3 weeks of December and have the national championship game the second weekend in Jan or the weekend before the Super Bowl.
 
The difficulty of schedule is taken into account. How do you figure that? Beating Bama in 2000 and beating Bama in 2009 are vastly different things. The margin of victory is going to say something about the quality of the teams involved. If Florida lays 50 points on GA, does that not speak to the ability of the Gators? Running up the score is not a sign of sportsmanship, but it is a sign of what the team is capable of doing.

As for having the players injured or having crappy weather, those things effect the outcome of games and no one cares. If Bama loses because 6 of our starters are injured, we don't get an asterisk by the game, we have a loss. If one team is pass oriented and one is run oriented, the run oriented team will win in a blizzard. That does not mean the pass oriented team gets any points for losing in a blizzard.
I don't have any problem with some use of a correction factor for difficulty of schedule and wins against stong opponents. Margin of victory should not be one of them. It's been done and it didn't work.
 
I don't have any problem with some use of a correction factor for difficulty of schedule and wins against stong opponents. Margin of victory should not be one of them. It's been done and it didn't work.

Hell none of it has actually worked. There has not been a single season in which there were not reasons for someone to complain.
 
Hell none of it has actually worked. There has not been a single season in which there were not reasons for someone to complain.
True but thats because you can't have a national championship by proxy. We need a playoff. We need to keep pestering our congress people till change occurs. I could really give a shit less anymore about the Rose Bowl/Orange Bowl, etc.

A national championship tournament would garner more attention and money then the BCS does.
 
Back
Top