Writers of Pro Football Prospectus 2008

Most Recent FO Features

HackenbergChr15.jpg

» SDA: From Death Valley to Happy Valley

While Saturday won't give us an answer to which is the best one-loss team, it is still a big week for conference races across the country.

12 Nov 2010

VN: FOA 2010 Projections in Review

by Bill Connelly

Thanks to Football Outsiders Almanac 2010 getting discounted 33 percent this week and my revisiting teams' five-year histories for yesterday's ESPN Insider piece, I think it's time to revisit the F/+ preseason projections. Which teams did we nail completely? On which teams did we completely whiff, and why? With plenty of both the regular season and postseason remaining, we cannot yet revisit FO's win projections, but we can look at which teams have varied significantly in terms of their F/+ rankings.

As FOA owners know, we expanded the reach of the projections this year to include not only FEI and S&P+ data, but also data about recruiting, coaching changes, and other change factors (returning starters/lettermen, late-season momentum, turnover margin, fumble recovery percentage, etc.). In all, the projections have not been too bad. Forty-seven of 120 FBS teams are within 10 spots of their projected rankings, and many which aren't are mid-major teams still reaping the benefits of a tough non-conference schedule without completely paying for their weak conference schedules (No. 71 Florida International, I'm looking in your general direction).

Primarily because of the five-year history factor described in yesterday's Insider article, we certainly missed on some of our top picks. Here was the projected F/+ Top Ten:

1. Florida (Current F/+ Rank: 23rd | Record: 6-3).
2. Alabama (Current F/+ Rank: fifth | Record: 7-2).
3. Texas (Current F/+ Rank: 54th | Record: 4-5).
4. Ohio State (Current F/+ rank: fourth | Record: 8-1).
5. Virginia Tech (Current F/+ rank: 13th | Record: 7-2).
6. Oklahoma (Current F/+ rank: 17th | Record: 7-2).
7. TCU (Current F/+ rank: second | Record: 10-0).
8. Penn State (Current F/+ rank: 47th | Record: 6-3).
9. Oregon (Current F/+ rank: 10th | Record: 9-0).
10. LSU (Current F/+ rank: 12th | Record: 8-1).

Really, that list includes three whiffs (Florida, Texas, and Penn State) and seven solid picks. Boise State should have been in the mix as well, instead of lingering at 15th, but in all, this isn't a terrible showing. (Nobody else had Florida and Texas tumbling to a great degree either.)

Recent history seemed to have less impact this season than it has in other recent years, and in the offseason, we will have to determine whether it will play as large a role in next year's projections. My guess is that it will. The 2007 season was a crazy one as well, and projections based around recent history wouldn't have done very well that season either, but 2008 would have been just fine in that regard. You don't tweak formulas due to outlier seasons, but again, those are decisions we will make this offseason.

I wanted to revisit some of the other "misses" this week to see if there are themes among them, or if they were just random surprises. I won't discuss the teams from yesterday's ESPN piece -- Stanford, Illinois, Mississippi State, Texas, BYU, Boston College -- in much detail because, well, to the extent that our projections were off on those teams, it was pretty clearly because of five-year history and its (lack of) impact. Below are some teams we missed, and not just because of recent history.

We Underestimated 'Em

N.C. State. FOA 2010 Projection: 63rd | Current F/+ Rank: 22nd | Record: 6-3. The Wolfpack rank 12th in Defensive F/+ ... just slightly off from our projection of 78th. They have been efficient against the run and have done a wonderful job of preventing big plays in the air. Star linebacker Nate Irving has returned to the lineup this year after a car accident sidelined him in 2009, and he has made a world of difference. He leads the team in tackles (65), tackles for loss (10.5, an excellent total for a linebacker), and quarterback hurries (eight), plus he's broken up four passes. He has combined with Audie Cole and Brandan Bishop to form one of the better, most underrated linebacker corps in the country.

Hawaii. FOA 2010 Projection: 89th | Current F/+ Rank: 38th | Record: 7-3. Hawaii's projections were weighed down by both a poor defensive outlook (105th in projected Defensive F/+) and strength of schedule (90th). They have already exceeded their projected 6.6 wins, and they have done so by vastly exceeding expectations on both sides of the ball. Quarterback Bryant Moniz has thrived in his second season in charge of the offense, and the skill position players around him -- running back Alex Green and especially receiver Greg Salas -- have been outstanding. Clearly the Rainbow Warriors weren't up for the task at Boise State last weekend, but who is?

Kansas State. FOA 2010 Projection: 72nd | Current F/+ Rank: 41st | Record: 6-3. Chalk this one up to recent history and the Bill Snyder effect. It appears that FO missed on teams with first- and second-year coaches more than anything else, and obviously a five-year history number (which include three seasons of Ron Prince and one from the first Snyder era) wasn't going to tell us an inordinate amount about how the 2010 Wildcats would play. It is still unclear to me what the Wildcats' ceiling might be under Snyder -- even with the Brown brothers coming to town, I don't see another conference-title run in their future -- but it is obvious that the ceiling is higher with Snyder than with Prince.

Missouri. FOA 2010 Projection: 41st | Current F/+ Rank: 14th | Record: 7-2. The Tigers are 3-1 against F/+ Top 40 teams, and that has helped them hold onto a Top 15 spot despite the upset loss to Texas Tech. As with N.C. State, the Tigers have exceeded projections on the defensive side of the ball (Projected Def. F/+ rank: 45th), where an experienced secondary and improved pass rush have improved the Tigers from 78th in Def. Passing S&P+ to fourth. Injuries are taking their toll on the Tigers -- they have slipped to 67th against the run, due in part to injuries to defensive tackle Dominique Hamilton and linebacker Luke Lambert. We will see if a defense that wore down against Baron Batch and Texas Tech can hold their form against Daniel Thomas and Kansas State in this weekend's F/+ Overachiever Bowl.

Others: Central Florida, Army, Florida International, Arkansas State, San Diego State, Nevada, SMU.

Pie In The Sky

When the projections were finalized, three teams stood out to me as worrisome picks. All three teams -- Kansas, Tennessee, and Texas Tech -- were breaking in new coaches and trying to replace a solid amount of talent, and it was going to be very difficult for them to live up to these expectations. None of the three have, and only Texas Tech (at 5-4) has managed a winning record to date. Coaching changes -- and the high variability and unpredictability they bring along -- are most likely always going to be a bugaboo for projections.

Kansas. FOA 2010 Projection: 51st | Current F/+ Rank: 115th | Record: 3-6. Among other things, Kansas is proof of the limitations of the "returning starters" figure. The Jayhawks returned six offensive starters from last year's squad, but three of the players they lost -- quarterback Todd Reesing and receivers Dez Briscoe and Kerry Meier -- were potentially more valuable than the other eight starters combined. When you combine misleading returning starter data with a first-year coach (affable, religious, "no cursing" Turner Gill seems to be the exact opposite of Mark Mangino in every way, for better or worse), this was a solid recipe for disaster. It is impressive that Kansas has managed three wins despite an overall F/+ rank of 115th and the fact that they lost to North Dakota State to start the season. They have pulled off one inexplicable upset (albeit to another team on this portion of the list, Georgia Tech), one inexplicable comeback (they were down 28 points to Colorado in the fourth quarter last Saturday and won; it was the final straw for Dan Hawkins' tenure in Boulder), and one win over the worst-ranked team in FBS (New Mexico State). With remaining games against Nebraska, Oklahoma State, and Missouri, let's just say their odds of getting to 4-8 are not high. Of course, I said the same about their odds of getting to three wins halfway through the Colorado game.

Tennessee. FOA 2010 Projection: 20th | Current F/+ Rank: 73rd | Record: 3-6. A first-year coach and misleading recruiting rankings -- quite a few high-ranked members of recent recruiting classes enrolled initially at Tennessee but are no longer with the team -- resulted in the FO bar getting set far too high for the Volunteers. Despite one of the stronger strengths of schedule around, the Vols have still slipped to 73rd, in part because they have no definitive strengths. They were projected in the Top 25 both offensively and defensively, but they currently rank 74th in Offensive F/+ (47th in Passing S&P+ but 78th in Passing Downs S&P+) and 65th in Defensive F/+ (67th in Passing Downs S&P+). They are young and thin, and with their schedule, they just haven't had much of a chance.

Purdue. FOA 2010 Projection: 46th | Current F/+ Rank: 96th | Record: 4-5. The Boilers were projected to take a step forward in Danny Hope's second season (coaches' second seasons are almost as variable as first seasons), but injuries have wrecked an offense that currently ranks 109th in Defensive F/+. Starting quarterback Robert Marve was lost for the season in his fourth game, receiver Keith Smith caught 18 passes in seven quarters, then tore up his knee. Star running back Ralph Bolden tore his ACL in the spring. The young Boilermakers have had to learn on the fly, and it hasn't been pretty. If everybody comes back healthy next season, Purdue could be in store for a solid bounce-back year. But with a slumping recent history, we probably won't be in position to get that one right either.

Georgia Tech. FOA 2010 Projection: 23rd | Current F/+ Rank: 57th | Record: 5-4. The Yellow Jackets have underachieved equally on both sides of the ball. An offense that was projected sixth in Off. F/+ ranks 38th, primarily because they miss Demaryius Thomas more than anybody could have imagined. The absence of the first receiver taken in last year's NFL Draft has been the catalyst behind Tech falling from 16th in Passing S&P+ to 108th. Thomas masked quarterback Josh Nesbitt's inaccuracy as much as anyone could, it seems. With no threat of the pass, the running game has found fewer open lanes. Meanwhile, Paul Johnson welcomed defensive coordinator Al Groh and his 3-4 system into the mix in 2010, and there has been more of a breaking-in period than expected. Georgia Tech ranked 64th in Defensive F/+ last season and were projected 46th. Instead, they have fallen to 82nd. They are equally bad against the run (90th in Def. Rushing S&P+) and the pass (84th in Def. Passing S&P+). Blitzing has been a strength, but clearly the personnel hasn't meshed well just yet, and they likely miss defensive end Derrick Morgan as well.

Others: Wake Forest, Connecticut, Minnesota, Penn State, Texas Tech.

F/+ Rankings

Full rankings here.

F/+ Top 25 (After Ten Weeks)
RK TEAM REC. F/+ LAST
WK
CHG S&P+ RK FEI RK OFF.
F/+
RK DEF.
F/+
RK
1 Boise State 8-0 +32.0% 1 +0 286.8 1 0.243 7 +19.0% 3 +13.0% 7
2 TCU 10-0 +28.3% 8 +6 270.6 3 0.249 4 +12.1% 14 +16.2% 1
3 Auburn 10-0 +28.2% 2 -1 256.4 5 0.314 1 +24.3% 1 +3.9% 38
4 Ohio State 8-1 +26.3% 6 +2 271.2 2 0.205 12 +13.3% 9 +13.0% 8
5 Alabama 7-2 +25.4% 5 +0 259.9 4 0.242 8 +16.3% 4 +9.1% 17
6 Stanford 8-1 +22.8% 12 +6 248.0 7 0.247 5 +14.4% 6 +8.3% 22
7 Iowa 7-2 +22.6% 4 -3 252.8 6 0.221 9 +7.2% 28 +15.4% 3
8 Nebraska 8-1 +21.9% 3 -5 244.6 13 0.247 5 +10.6% 19 +11.3% 11
9 Arkansas 7-2 +20.8% 24 +15 246.3 9 0.217 11 +14.9% 5 +5.9% 30
10 Oregon 9-0 +20.8% 10 +0 235.9 20 0.267 2 +12.2% 13 +8.6% 21
F/+ Top 25 (After Ten Weeks)
RK TEAM REC. F/+ LAST
WK
CHG S&P+ RK FEI RK OFF.
F/+
RK DEF.
F/+
RK
11 Miami-FL 6-3 +20.4% 16 +5 247.9 8 0.201 14 +4.9% 36 +15.5% 2
12 LSU 8-1 +20.2% 21 +9 234.0 24 0.264 3 +8.4% 25 +11.8% 10
13 Virginia Tech 7-2 +20.1% 9 -4 242.5 15 0.220 10 +12.0% 15 +8.0% 23
14 Missouri 7-2 +19.9% 7 -7 244.9 12 0.204 13 +9.9% 22 +9.9% 13
15 Oklahoma State 8-1 +18.9% 19 +4 245.7 10 0.182 17 +12.5% 11 +6.4% 27
16 Wisconsin 8-1 +17.4% 13 -3 237.5 17 0.190 15 +14.1% 7 +3.2% 43
17 Oklahoma 7-2 +17.1% 14 -3 238.8 16 0.179 18 +9.5% 23 +7.6% 24
18 South Carolina 6-3 +17.0% 11 -7 245.5 11 0.144 23 +11.7% 17 +5.3% 32
19 Oregon State 4-4 +16.4% 15 -4 243.9 14 0.140 25 +13.4% 8 +3.0% 45
20 Pittsburgh 5-3 +15.9% 20 +0 235.9 19 0.169 19 +11.2% 18 +4.7% 36
F/+ Top 25 (After Ten Weeks)
RK TEAM REC. F/+ LAST
WK
CHG S&P+ RK FEI RK OFF.
F/+
RK DEF.
F/+
RK
21 Michigan State 9-1 +14.8% 17 -4 235.0 22 0.151 22 +6.2% 33 +8.6% 20
22 N.C. State 6-3 +13.9% 25 +3 223.4 34 0.190 15 +3.9% 41 +10.0% 12
23 Florida 6-3 +13.1% 30 +7 233.9 25 0.123 28 +5.6% 34 +7.5% 25
24 Illinois 5-4 +13.0% 27 +3 235.7 21 0.111 32 -1.0% 66 +13.9% 4
25 USC 6-3 +12.7% 22 -3 228.9 30 0.140 25 +12.2% 12 +0.5% 53

Boise State and TCU may currently be on the outside looking in when it comes to the BCS standings, but the two mid-majors have surged to the top of the F/+ rankings. Our rankings are designed not to automatically punish a team for playing a weak schedule -- you can maintain high rankings if you decimate your schedule as well as BCS conference powers would. With Boise State's consistency and TCU's incredible performance against Utah, I think it is safe to say that these teams have performed as well as Auburn, Oregon or anybody else would have with the schedule they were given. And unlike Oregon, they haven't waited until the 50-play mark to get rolling.

(The Ducks, by the way, face an interesting landmine game this weekend in Berkeley. California has been the most high-variability team in the country this year, and they have hinted at an upside solid enough to knock Oregon out. Unfortunately for the Golden Bears, it seems like it has been a while since they played at a high level against a good team.)

"What The ...?" Team of the Week

Arkansas. Aside from TCU-Utah, the Razorbacks put together the most impressive performance of the week, pummeling a good South Carolina team on the road. Others don't think quite as highly of South Carolina as our numbers do, but this was a great performance no matter how you slice it. The Hogs have now held a fourth-quarter lead in every single game of the season. With a remaining schedule of UTEP, Mississippi State, and LSU, they will have an outside chance at a 10-2 year.

AP + F/+ = BCS?

Here's your weekly look at how the BCS rankings would look if determined from a 60-40 split of AP Poll and F/+ rankings. It really is impressive that none of the top four teams hold the same spot in both this list and the actual BCS standings. In some years, the "Who should play in the BCS title game?" answer is obvious. In other years, it isn't even close.

1. Auburn (10-0)
2. TCU (10-0)
3. Boise State (8-0)
4. Oregon (9-0)
5. Ohio State (8-1)
6. Stanford (8-1)
7. LSU (8-1)
8. Nebraska (8-1)
9. Alabama (7-2)
10. Wisconsin (8-1)

Upset Watch

Ole Miss over Tennessee. Spread: Ole Miss +2.5 | F/+ Projection: Ole Miss by 4.0. For all the numbers know, Jeremiah Masoli is healthy and playing. If he's not either of those things (he suffered a concussion last week and is questionable), then Tennessee should be able to stop the Rebels enough to counteract their own offensive issues.

South Carolina over Florida. Spread: South Carolina +6.5 | F/+ Projection: South Carolina by 0.4. Considering the game is in Gainesville, this pick surprised me a bit. The numbers are not giving any extra credit to Florida for picking up the pace offensively in recent weeks, so that might account for some of the surprise. Florida has one of the best standard downs offenses in the country, and the Gamecocks will have to figure out how to create more success for running back Marcus Lattimore than his 11-carry, 30-yard performance of a week ago.

Notre Dame over Utah. Spread: Utah -5.5 | F/+ Projection: Utah by 2.0. The numbers say one thing here, but none of it really matters. For very different reasons, these two teams have had a tough couple of weeks between the ears. The Notre Dame family is still dealing with the grief involved with Declan Sullivan's death (their initial on-field response was a blown lead and a loss to Tulsa), and nobody would really blame them if they struggled the rest of the season. Meanwhile, Utah got their doors blown off by TCU, and they are all but certainly dealing with disappointment and a bit of embarrassment. There is no way of knowing how either team will respond.

Duke over Boston College. Spread: Boston College -3.5 | F/+ Projection: Boston College by 1.9. Like Florida, BC has performed better in recent weeks. But with an offense that is still mostly dead (they scored only 23 against Wake Forest despite five Wake turnovers), the Eagles might let Duke and their decent passing game hang around a while.

The Playlist

In honor of some of the FOA 2010 projections ...

"Am I Right or Wrong," by Son House
"Am I Wrong," by Keb' Mo'
"The Cow's Wrong," by The Beta Band
"Everybody's Wrong," by Buffalo Springfield
"I Might Be Wrong," by Radiohead
"I Was Wrong," by Social Distortion
"Something's Always Wrong," by Toad The Wet Sprocket
"What Am I Doing Wrong?" by David Gray
"Wrong 'Em Boyo," by The Clash
"The Wrong Way," by TV On The Radio

Not sure which is the more enjoyable segway there -- Social Distortion to Toad The Wet Sprocket, or David Gray to The Clash.

Closing Thoughts

I like the WAC's addition of Texas State and UT-San Antonio to their future roster quite a bit. The schools come from football-fertile areas, and given a few years to grow, they should be able to compete in the weakened WAC. It was a solid move on their part. Still, it had to sting when Montana considered a possible WAC invitation and said, "Nah, I think we'll stick to the Big Sky instead." It might have made business sense for Montana to do so, but it certainly did nothing for the WAC's image in the process.

Posted by: Bill Connelly on 12 Nov 2010

17 comments, Last at 13 Nov 2010, 10:58am by Dennis

Comments

1
by cfn_ms :: Fri, 11/12/2010 - 12:51pm

is what was different about 2007 and 2010 to make them so wacky? Was there anything unusual in the preseason data that in retrospect would indicate that some level of chaos was coming? If there was, then that's something that could be very useful in guessing when the next wacky season is coming.

2
by Eddo :: Fri, 11/12/2010 - 1:31pm

I feel like 2007 was much more wacky than this year. The major players for the title game were ranked 3rd, 6th, 11th, and 22nd in the preseason AP poll. The #1 team, Alabama, losing to a solid South Carolina team and a very good LSU team (on the road) shouldn't be too surprising. Ditto for the #2 team, Ohio State, losing to a very good Wisconsin team (in Madison).

Auburn's ascension is unexpected, but not that surprising. Any undefeated SEC team will be ranked in the top four come this time of year, and it just so happens that it's Auburn this year instead of Alabama or LSU.

Oregon was already considered to be the most likely Pac-10 champion, and has been a top-ten-level team for most of the decade.

In 2007, however, you had Missouri and Kansas playing for the #1 ranking in November. USC had lost to Stanford. Michigan lost to Appalachian State. Illinois came out of nowhere to finish second in the Big Ten and go to the Rose Bowl. Boston College and South Florida each rose to #2 at one point. Only two major-conference teams entered the bowls with fewer than two losses (Ohio State and Kansas). The eventual BCS champion was ranked 7th in the penultimate BCS standings.

This year, the top teams have all been upset by much more high-quality foes. The teams at the top aren't the usual suspects, necessarily, but consist of an every-few-years SEC power (Auburn), a consistent BCS bowl challenger (Oregon), and two recent non-AQ powers that were top-six teams in the preseason (Boise State and TCU). There's almost no chance a two-loss team makes the title game, let alone winds up the BCS champ.

This year's been awesome, in that we have a good shot at a non-AQ team reaching the title game, but I don't see it as being nearly as crazy as 2007. Last year's wire-to-wire act by Alabama and Texas probably makes it seem like this year is pretty crazy, but I actually think 2008 was a bit more (two one-loss teams in the title game, with the weird three-way tie atop the Big XII standings and an undefeated Utah on the outside looking in).

------

ADDENDUM: cfn_ms, you definitely know more about college football than I do, so if I've gotten any facts wrong or if I'm missing something that made 2007 so wacky, please forgive me (and correct me, as I'd like to know your point of view and reasons).

4
by cfn_ms :: Fri, 11/12/2010 - 2:46pm

I would personally agree that 2007 was far crazier than this year, and unless things get really wacky the rest of the way in 2010, then that year will really stand out as the craziest I can remember by a lot. I based my equivalence comment on Bill's note (i.e. while I feel that 2007 was craziest, I haven't thought deeply about it and could well be wrong):


Recent history seemed to have less impact this season than it has in other recent years, and in the offseason, we will have to determine whether it will play as large a role in next year's projections. My guess is that it will. The 2007 season was a crazy one as well, and projections based around recent history wouldn't have done very well that season either, but 2008 would have been just fine in that regard. You don't tweak formulas due to outlier seasons, but again, those are decisions we will make this offseason.

3
by Drew (not verified) :: Fri, 11/12/2010 - 1:53pm

The fundamental problem with this article is that it asks whether your statistical model does a good job of predicting which teams will do well in your statistical model, rather than whether your statistical model does a good job predicing which teams will win games.

6
by cfn_ms :: Fri, 11/12/2010 - 3:03pm

I think that, with a number of games remaining, it's fair to wait and see on the W/L projections as noted:

With plenty of both the regular season and postseason remaining, we cannot yet revisit FO's win projections, but we can look at which teams have varied significantly in terms of their F/+ rankings.

That said, it might be interesting to see how well F /+ did in picking over/unders for teams that have clinched one way or the other at least.

I'd also think that it's very much worth revisiting at the end of the regular season; I'm pretty sure that FOA's win projections didn't include CCG's or bowls, so there's really no need to wait that long to do the analysis.

5
by Eddo :: Fri, 11/12/2010 - 2:52pm

Kind of. A good model should also correlate well with itself moving forward, as well as with wins.

10
by Bill Connelly :: Fri, 11/12/2010 - 4:46pm

Really, you probably want to be able to do either (or both) of two things if you're using math to talk about sports -- evaluate and predict. Evaluating was my initial goal, and it stands to reason that if we can guess how teams will perform in a given model, and I like how that model evaluates teams, then there is value there. I love the results the F/+ model produces; I love how it irons out quirks in both S&P+ and FEI, and as a whole, I tend to agree with its judgments more than any other computer model. So if we can predict how a nice model will play out, that's valuable.

Making predictions based off of that model, however, is a completely different can of worms. That might require heavier weighting of more recent games or other factors that pure F/+ does not (and should not, as a full-season evaluative tool) take into account. It's clear that my first stab at predictions based on F/+ (which is basically the same method I used last year with S&P+, only with F/+) has not been tremendously successful, at least not as it pertains to the spread. That doesn't impact its evaluative ability -- it just means we haven't yet figured out the best way to use it as a predictive tool.

7
by IB (not verified) :: Fri, 11/12/2010 - 3:25pm

Six of your top ten teams have lost two games or more with a few weeks left in the season. That's not particularly accurate. Also, F+ is not performing particularly well as a predictive tool (it's picking winners against the spread at something like a 45% rate right now, at least in the games chosen in the other article).

I've never been particularly enamored by the college football models FO employs, because a lot of it is just some fancy math on top of a foundation of "teams that have been good will still be good". That's a decent heuristic for college football, but there's nothing all that complicated about it; I could list the same teams year after year in the top 25, and I'll get 12-15 of them every time.

But, as a model, that leaves a lot to be desired. "Who could have figured Florida and Texas would do so badly" is a fair point, though there were some signs that Florida would struggle even before the season. But FO has gone into the season with a lot of indefensible projections, like Tennessee and Purdue this year, merely on the basis that those teams have previously been good (see also Boston College, pretty much every season), as though the mere fact of being Tennessee will make the team better than it is.

Once again, good teams stay good in college football might be a useful heuristic. But if that's the extent to which this model has advanced, then the model isn't worth very much right now.

8
by Bill Connelly :: Fri, 11/12/2010 - 3:50pm

"...it is just some fancy math on top of a foundation of 'teams that have been good will still be good'."

That foundation came from the math. We didn't start with that idea and pile math on top of it. To the extent that it is a heuristic, it is that way because the numbers made it so.

9
by Kal :: Fri, 11/12/2010 - 4:45pm

To be fair, Tennessee's prediction was based on a lot of traditional ones:

1) new coach, which tends to benefit the team
2) good recruiting class, mentioned above
3) good overall program; good programs don't tend to slide significantly for very long
4) number of returning starters

All of those things make sense as a correlation for success in future years. Oregon, for instance, had basically all of those things as well and was (correctly) predicted to do well. Simply put, sometimes you'll get it wrong.

That being said F/+ doesn't seem to be doing a good job as a predictor or as a mitigator of the other's successes and failures. I think part of that is that the scale is really off; the difference between teams in FEI is much smaller than the difference between teams in S&P, and that throws things quite off. FEI does seem to be doing fairly well on predictive qualities though (based on the bcstoys link). I'm curious how S&P is doing by itself on prediction?

11
by Bill Connelly :: Fri, 11/12/2010 - 4:47pm

Unfortunately, I'll have to wait until the offseason to catch up on S&P+ picks -- doing the F/+ picks on a week-to-week basis, I haven't had time to do it both ways.

12
by Thok :: Fri, 11/12/2010 - 6:27pm

California has been the most high-variability team in the country this year, and they have hinted at an upside solid enough to knock Oregon out.

I would argue that Cal has issues with athletic teams (basically if you don't have a certain amount of athleticism Cal crushes you, but above the critical amount you crush Cal, with Arizona having the rare balance to play evenly with Cal), which bodes poorly for them against Oregon even before you take into account the fact that Mansion/Sweeney are much, much worse than Riley and that's saying something given how Riley was playing.

(Seriously, Mansion was playing at Joe Ayoob levels last week, which means Cal should honestly start playing tight ends at quarterback.)

13
by cfn_ms :: Fri, 11/12/2010 - 7:00pm

Or just go with Vereen in the Wildcat.

14
by zlionsfan :: Sat, 11/13/2010 - 12:38am

The Big Sky is a pretty strong I-AA football conference, plus it's expanding as well ... although it's getting to the point that some schools likely will move on (14 is really too much for a conference at any level).

Strangely, I've seen Portland State's name dropped with respect to possible WAC membership. Yes, Portland's a relatively big market, but PSU is a mostly-commuter school, I believe (they offer an all-online master's of software engineering that I'm working on), and their athletic budget isn't anywhere near what a WAC school spends. Also, the major teams have struggled lately ... would they draw talent similar to an Oregon State or a Washington if they were a I-A school? Don't know.

15
by Thok :: Sat, 11/13/2010 - 9:35am

I found some 2007 athletic budgets and Portland St's isn't that out of line. It listed them as $9 million, when $18 million would be middle of the pack of the WAC in 2007 and $27 million would be comparable to the top of the WAC in 2007 (Boise St and Hawaii, and Hawaii has unique issues with travel costs that forces them to have a high budget for the WAC.)

It would be a jump, but not as much as you'd think.

16
by Bill Connelly :: Sat, 11/13/2010 - 9:42am

I actually thought it was a very smart move on Montana's part. The Big Sky IS a strong conference, and moving up to FBS is often a very foolish move. But without Boise, Hawaii and Fresno, the WAC is fighting a very well-earned perceptions problem, and Montana passing on them doesn't help that.

17
by Dennis :: Sat, 11/13/2010 - 10:58am

The WAC is in big trouble obviously. I don't think Montana not wanting to move up is that big a deal. The obvious problem is they've lost their top teams. They just announce that the University of Denver is joining as a non-football member, which will give them some help in basketball with the market. But they are pretty much dead as a football conference.