FanPost

ACC versus the Computers week 6

First a warning this is going to be a long post as I will cover the computers themselves. Last week was pretty good for the ACC. In conference play the only non optimal but not bad result was VT over GT. In the short run it does not really matter but in the long run there was more upside with GT. In ACC OCC games last week only UNC let us down the slight bright side ECU was ranked slightly higher going into the game in the computers so not a huge negative. In pure OOC games a lot of good happened. WVU beat Okie St insuring ACC stays ahead and by a significant margin the B12. UGA eliminated the only non Pac team that can finish higher in the polls then the Clemson/FSU winner if they go undefeated. USCe beat UCF which all but ensures that even a 1 loss FSU/Clemson finishes ahead of undefeated UofL in the computers. USCe, UF, and UGA can almost be seen as minor ACC teams this year because of they each play 2 games against ACC teams.

First up a little bit of know the polls. I wanted to go more in depth but the computer already ate my first attempt. The computers make up 1/3 of the BCS formula. The computer section of the BCS formula is made up of 6 computers where for every team their top and bottom ranking are thrown out so its an average of the 4 middle polls. Some terminology real quick I will be using. OR means Opponents record which is what is your Opponents WP. OO means Opponents Opponents or for FSU for example who did BC play. WP is win percentage which is what it says what is your record as a percentage. RPI is the formula used in NCAA basketball it is WP*(OR+OO).

First up is Billingsley mostly because it is stupid and a very bad poll. How bad? Historically it has been the poll thrown out more then all other polls combined. And this year it will likely be thrown out in even more often then its historic average. I could honestly see this poll being thrown out for every AQ school in the final BCS poll.

Why is it so bad? He is the only computer to use preseason rankings (Sargarin uses preseason until week 5 or 7 then drops that completely this is done so it looks right at the beginning of the year and really the only way to do it with his method this early). The games are rated on the week they are played and not corrected for later results. The first 4 weeks are heavily weighted. The reason these are bad methodology this year is last year the SEC basically was about 3 OOC wins from the perfect season for computer algorithm season. This means in his preseason poll SEC teams where very high. Last year the B12 was playing at near historic levels for the conference again meaning higher preseason rankings. The ACC had a historic bad year meaning lower preseason rankings. The B1G was down from even its recent levels meaning lower preseason rankings. And this year the Pac is playing close to historic highs so far which they where not last year. What this basically means is based on the information from the first 5 weeks the teams are not in general ranked that way in his formula because its based on last years numbers. In his poll if say LSU falls apart and loses 5 this season for the rest of the year UGA gets credit in the final poll for beating a top 5 team despite LSU really being a 35th or worse team. But like I said it will be the dropped poll for just about every team that matters so don't worry about this one.

Next up is the most famous of the computer polls with Sagarin. He uses a method called Elo which is used in chess. Wikipedia has a good article about the exact math but the big thing about his method is its more of a Team A beats Team B which Beat Team C so Team A is better then Team B. Of course there is more to it then that but that is the big thing to take away from his methodology. His SOS is derived from his rankings and includes adjustment for game location.

Wolfe is the most secretive but what is known its similar to ELO. It basically ranks team based on what their chances of winning against other teams are. He does this by looking at common opponents. He not only ranks FCS but ranks every school that plays college football. I am fairly sure he includes game location.

The next 3 polls all use some variation of RPI which again is WP*(OR+OO). The weighting is different with each method so will come up with different exact results but at the end of the day this is the basic formula here. None of these use a preseason ranking. Unlike Sagarin these polls don't recognize Team A beat Team B beat Team C so Team A is better then Team C. Instead the way it sees it is Team A "RPI" is higher then B "RPI" which may or may not be higher then team C RPI.

Anderson and Hester twist on the formula is adding a conference SOS to the mix. I don't personally like this because for polls like this there really is no such thing as the ACC. There is FSU ACC which does not include playing FSU or any Coastal team except Miami and Pitt. While Clemson's ACC does not include playing Pitt and Miami. For the best teams in a conference this gives them a bonus while punishing the bottom. Game location also is factored into the mix. They also give a slight bonus to beating top teams while punishing losses to bottom teams. This year though the Pac is the clear number 1 conference for this poll or will be once it is released and right now the SEC and ACC are about equal in second currently.

Colley Matrix twist first he includes BCS teams but puts them in blocks based on how good they are. It also ignores game location. He also goes many OO deep for his SOS rankings but the more often he adds an OO (think OOO OOOO OOOOO) the less weight that level has.

Massey twist he weighs later games slightly more then early games. He does care about game location. He does not include FCS. His is very simple.

In all of the ones I have done so far and will do in the future I basically ignore Billingsley because like I said above this year especially I just don't see his poll being used by any of the teams that matter. So onto this weeks games.

Ball State @ UVA Ball State is actually higher in the computers then UVA. UVA needs to win but not a horrible loss if they don't.

Maryland @ FSU Yeah FSU needs to win.

UNC @ VT VT needs to win. UNC winning this game would be fairly bad.

Army @ BC I hope BC does not have a let down because they should and need to win this game as it would be a very bad loss.

NCST @ WF WF winning any conference game except maybe UNC would be bad bad bad

GT @ Miami Miami needs to win this one. Its not a horrable loss but Miami winning maximizes ACC teams towards the top and gives FSU a big boost when we play.

Clemson @ Cuse Cuse is allowed to win a couple of ACC games but this is not one of them.

OOC games of note

WKU @ Louisiana-Monroe believe it or not this has some effect. LM wins and WF should go ahead of UK in the polls for the rest of the year while WKU winning UK goes ahead of WF in the polls this matters for the OO portion.

ASU @ ND ND is our friend this year and this can help close the gap a bit with the Pac.

tOSU @ NW NW knocking off tOSU will make Cuse look a little better but more importantly take tOSU completely out of picture.

Notable game of the week that has only minor effect on the ACC and FSU is UW @ Stanford. Either of these teams wins out they will be above FSU. That said Stanford winning means ND could beat them and Pitt could beat ND and get a boost that way. Not likely but this is the biggest game of the week just wanted to give the ACC perspective.

SB Nation Featured Video
X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Tomahawk Nation

You must be a member of Tomahawk Nation to participate.

We have our own Community Guidelines at Tomahawk Nation. You should read them.

Join Tomahawk Nation

You must be a member of Tomahawk Nation to participate.

We have our own Community Guidelines at Tomahawk Nation. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9341_tracker