Week 9 Game Threads

Moderators: studcrackers, bleu

User avatar
Los Soles
Sixth Man
Posts: 1,858
And1: 368
Joined: Sep 29, 2010
Location: Boise, ID
     

Re: Week 9 Game Threads 

Post#61 » by Los Soles » Tue Nov 1, 2016 12:14 pm

Yeah, so...it doesn't actually matter. We agree. :wink:

I also agree with this: through 3 teams, the Big 10 wins. CFP committee will show that. Fair enough. And if I'm an Ohio State fan, I'm thrilled where things stand after this past weekend, because Ohio State is now one of maybe 5 or 6 teams that controls its own destiny: win out and you're in. (Caveat: I'm not so thrilled with how Ohio State looked last weekend –- haven't looked at all like a playoff team for a couple weeks now.)

But 4 top-10 teams is a joke: Nebraska is not a top-10 team. Voters are stupid. Nebraska is still getting extra credit for barely surviving a *ranked* Oregon (at home!)...that's now 1-4 in the Pac-12. Voters aren't capable of processing much more than record, so they see a 1-loss power-5 team and fail to notice that their best win was, what, Northwestern? Or the 8th best team in the Pac-12? Now, if Nebraska beats Ohio State, then Nebraska is legit in the top 10...but then Ohio State would no longer be a top-10 team. So I don't think the CFP committee will be as kind to Nebraska, and then I don't think the 4-in-the-top-10 thing will last in any ranking system.

I think that, conference as a whole, the Pac 12 is better than the Big 10. AP/coaches polls don't assess this; Sagarin, FPI, etc. do. And they say the Pac-12 is better, and it's not really all that close. Better in the middle, better by simple average, less garbage at the bottom, etc. And the middle matters: it's not just about the top 2–3 teams (see: Ohio State vs Penn State :o). We don't yet know who has the better #1, but it does seem like there's a pretty good chance both Washington and the Ohio State/Michigan winner will be in the playoff, so maybe we'll get to see that one settled on the field. Classic Rose Bowl #2 vs #3 matchup?!?
Goober
GOTB's Cancun
Posts: 9,615
And1: 3,621
Joined: Jun 09, 2014
     

Re: Week 9 Game Threads 

Post#62 » by Goober » Tue Nov 1, 2016 1:43 pm

Nebraska almost beat Wisconsin in wisconsin. They looked like a top 10 team and are deserving of their ranking
Image
User avatar
Los Soles
Sixth Man
Posts: 1,858
And1: 368
Joined: Sep 29, 2010
Location: Boise, ID
     

Re: Week 9 Game Threads 

Post#63 » by Los Soles » Tue Nov 1, 2016 11:07 pm

Goober wrote:Nebraska almost beat Wisconsin in wisconsin. They looked like a top 10 team and are deserving of their ranking

Nebraska in the rankings...

Polls

  • AP: 9
  • Coaches: 10
Computers

  • Sagarin: 23
  • FPI: 28
  • Massey: 18
Teams ranked behind Nebraska in AP, but ahead of Nebraska in all of Sagarin, FPI, and Massey:

  • Auburn
  • LSU
  • Colorado
  • USC
  • Stanford
  • FSU
  • WSU
  • Florida
I don't think it's a black-and-white, computers are always better than the polls. But the computers are capable of consuming and processing a lot more information. When all the computers agree to something that significantly diverges from your opinion (and/or popular opinion), you might consider why that might be.

Similar example: a week ago, Baylor and West Virginia were undefeated, and therefore top-10 teams in both the Coaches and AP poll. FPI said, "I see that they're undefeated, but their SOS and point-differential and home-away splits and common opponent comparisons lead me to believe they are not top-10 teams." Lo and behold, one week later...
User avatar
Arda K
RealGM
Posts: 47,969
And1: 18,756
Joined: Apr 11, 2007
Location: Frisco
     

Re: Week 9 Game Threads 

Post#64 » by Arda K » Wed Nov 2, 2016 2:23 am

You don't seem to understand how computer rankings work, they are totally dependent of having large sample size. Using them in a 13 game college football season is fatally flawed and using them midseason after 8 games is totally worthless. That's why you see silly things like 3-5 team ranked 16th on sagarin.

Sagarin works better during March madness time since college basketball plays 3x games, but using it as a proof of anything at this juncture is a sign of not understanding rudimentary statistical concepts.
TheLowlySquire wrote:Wow, Arda! Huge!


Howard Mass wrote:Arda is not a terrorist. Arda is a good person.
User avatar
Los Soles
Sixth Man
Posts: 1,858
And1: 368
Joined: Sep 29, 2010
Location: Boise, ID
     

Re: Week 9 Game Threads 

Post#65 » by Los Soles » Wed Nov 2, 2016 6:49 pm

I didn't say anything about "proof" of anything. I think computer rankings are an invaluable tool to help us assess that which is extremely difficult to assess.

Small sample size is a fundamental problem of NCAA football rankings, not computer rankings in particular. Of course a program with more data is better than a program with less data, but we have what we have, so we do our best with it. Any attempt to rank NCAA football teams suffers from a lack of information. E.g., there have only been 5 games between the Pac-12 and Big-10, and none of those were #1 vs #1, #4 vs #4, etc., so how do we even begin to compare these conferences? We know Pac-12 #1 Washington crushed Big-10 cellar dweller Rutgers...so what? We know Big-10 #4 barely survived Pac-12 #8...but how much can we conclude about relative conference strength from that? As we strip it down, there's just not much info from those five games. And any one game could be an anomaly/outlier; all teams have convoluted data because they have good days and bad days. (Is Ohio State the team that beat Oklahoma and Wisco on the road, or the team that lost to Penn State and barely survived Northwestern at home? Yes.)

Using computers allows us to incorporate more information than not using computers. The human mind, even one deeply immersed in college football, is only able to process a tiny bit of the information available, e.g., records; name brands; memories of a tiny percentage of the ~1000 D1 games that have been played. Computers can incorporate the information from literally all of those games.

When people look at a team like Nebraska, we look at their record and the scores of the games they played, and then if we're really digging deep, we can look at the schedules/records of the teams they played. And that's pretty much all we can incorporate mentally. But a well-crafted computer program incorporates the records and point-differentials of the teams Nebraska played...and the teams those teams played...and the teams those teams played...and the teams those teams played...and the teams those teams played...until we've incorporated all the information that there is. A computer program is literally looking at all ~1000 of those games to assign Nebraska a rating.

Ultimately, the dearth of data is a fantastic argument for...expanding the playoff. :thumbsup:
User avatar
Arda K
RealGM
Posts: 47,969
And1: 18,756
Joined: Apr 11, 2007
Location: Frisco
     

Re: Week 9 Game Threads 

Post#66 » by Arda K » Wed Nov 2, 2016 7:45 pm

yeah expand it to 16 so that 3-5 mississippi can make it to playoffs as they are 16th in sagarin :lol:
TheLowlySquire wrote:Wow, Arda! Huge!


Howard Mass wrote:Arda is not a terrorist. Arda is a good person.
User avatar
Los Soles
Sixth Man
Posts: 1,858
And1: 368
Joined: Sep 29, 2010
Location: Boise, ID
     

Re: Week 9 Game Threads 

Post#67 » by Los Soles » Wed Nov 2, 2016 8:15 pm

I've been saying 8: five conference champions, and then there's still space for, I dunno...Texas A&M? Ohio State? Louisville?

Sounds pretty much perfect to me.
User avatar
Latrell
Retired Mod
Retired Mod
Posts: 7,069
And1: 4,456
Joined: May 06, 2004
Location: Tuscaloosa
     

Re: Week 9 Game Threads 

Post#68 » by Latrell » Thu Nov 3, 2016 9:04 pm

8 is too many. I'll say 6 is fine with teams getting byes but let's just stick with 4 for a while. I mean we've gone 100+ years without a playoff system at all so let the current system breathe a little.
Image
User avatar
bwgood77
Global Mod
Global Mod
Posts: 33,741
And1: 13,732
Joined: Feb 06, 2009
Location: Austin
Contact:
   

Re: Week 9 Game Threads 

Post#69 » by bwgood77 » Sat Nov 5, 2016 6:31 am

Arda K wrote:yeah expand it to 16 so that 3-5 mississippi can make it to playoffs as they are 16th in sagarin :lol:


16 would be great if you like watching win and move on, lose and go home games more than meaningless bowl games, but to each his own.

But I love the college basketball tourney and would love a similar system cut by a couple of rounds. Leaves almost all fans of teams interested in much of the season for their team because they have some sort of chance to get in. You mention a losing team getting in, and I doubt that would happen, but at least it would be an easy win for a 1 seed.

I wish they just had the balls to play some playoff games and risk their great teams potentially losing.
User avatar
Arda K
RealGM
Posts: 47,969
And1: 18,756
Joined: Apr 11, 2007
Location: Frisco
     

Re: Week 9 Game Threads 

Post#70 » by Arda K » Sat Nov 5, 2016 10:40 am

uhh if your team is good enough just about every game is win or go home, the more you expand the playoff actually less important games there are during the season. f.e couple seasons ago losing to penn st more or less would have ended ohio st's chances at playing in the title, now that there are 4 playoff teams that game doesnt even matter ohio state still gets in if they win out. now expand that to 16 then half the games good teams play during the season essentially becomes meaningless.

i'd be fine with conference champions plus one or two wild cards but no more. last thing we need is to dilute the entire season so that some team that half assed their season to a 8-4 record gets hot or takes advantage of injuries to knock off a great team that worked their asses off for an entire season.
TheLowlySquire wrote:Wow, Arda! Huge!


Howard Mass wrote:Arda is not a terrorist. Arda is a good person.
User avatar
Arda K
RealGM
Posts: 47,969
And1: 18,756
Joined: Apr 11, 2007
Location: Frisco
     

Re: Week 9 Game Threads 

Post#71 » by Arda K » Sat Nov 5, 2016 5:45 pm

Los Soles wrote:in no case will there be "5 big ten teams ranked higher than second best pac 12 team".


let me help you count

College Football Playoff Rankings
RK TEAM REC TREND
1 Alabama 8-0 NR
2 Clemson 8-0 NR
3 Michigan 8-0 NR
4 Texas A&M 7-1 NR
5 Washington 8-0 NR
6 Ohio State 7-1 NR
7 Louisville 7-1 NR
8 Wisconsin 6-2 NR
9 Auburn 6-2 NR
10 Nebraska 7-1 NR
11 Florida 6-1 NR
12 Penn State 6-2 NR
13 LSU 5-2 NR
14 Oklahoma 6-2 NR
15 Colorado 6-2 NR
16 Utah 7-2 NR


:thinking:
TheLowlySquire wrote:Wow, Arda! Huge!


Howard Mass wrote:Arda is not a terrorist. Arda is a good person.
User avatar
Los Soles
Sixth Man
Posts: 1,858
And1: 368
Joined: Sep 29, 2010
Location: Boise, ID
     

Re: Week 9 Game Threads 

Post#72 » by Los Soles » Mon Nov 7, 2016 6:31 pm

Yeah, you were right on that. I was pretty surprised: I didn't think the CFP committee would diverge that much from AP, Coaches, Massey, Sagarin, and FPI. I thought they'd be closer to a blend between AP/Coaches and the computers, but they just kinda did their own thing. But just FYI, week 9 CFP ranking still doesn't actually matter, because there's still plenty of time for overrated teams to get exposed. I feel like it's already fairly obvious that those 1st rankings were pretty ridiculous (see: "Texas A&M").

One week ago, we were having a spirited debate about whether Nebraska was a top-10 team. Me and the computers said, "not even close." Refresh on Nebraska in the rankings...

Polls

  • AP: 21
  • Coaches: 20
Computers

  • Sagarin: 29
  • FPI: 37
  • Massey: 23
So, Nebraska dropped double-digits spots in the AP and Coaches. Did they get that much worse at football in a week...or were the computers right all along?? :wink:
User avatar
Arda K
RealGM
Posts: 47,969
And1: 18,756
Joined: Apr 11, 2007
Location: Frisco
     

Re: Week 9 Game Threads 

Post#73 » by Arda K » Tue Nov 8, 2016 11:59 am

sagarin still has usc 14th and ole miss 20th, until those are fixed computer rankings are garbage.
TheLowlySquire wrote:Wow, Arda! Huge!


Howard Mass wrote:Arda is not a terrorist. Arda is a good person.

Return to NCAA Football