The reason it works is because since 2002, roughly two preseason top 10 teams per year have ended the season unranked. The coaches have only avoided having top 10 teams fall out of the polls twice (2004 and 2007). All preseason polls are shaky, but the Coaches' Poll is the worst of them all. Coaches rank on reputation more than any other pollsters, and reputation factors in more in preseason polls than any of the rest of them.
Keep in mind that this list here doesn't necessarily predict that any team will finish the season unranked. It's merely the order of who I think is most to least likely to end the year outside the Coaches' Poll. When I list "potential losses" for each, it means that's the number of tossups (or worse) on the schedule in a worst case scenario. The magic number to hit is five, as about the only way to finish in the top 25 with five losses is to lose to the national champion twice like South Carolina did last year.
This is also not a ranking of overall quality of the teams. It's more about who has the lowest floor.
1. Oklahoma State
Where to start? Last year's 11-win season was the first in school history and only the fourth campaign to hit at least 10 wins ever. They had a luck rating of +0.95 in terms of Pythagorean wins, had a favorable +12 turnover margin, and lost just five starts to injury. Nearly everything that could have gone right did except for the six-point loss to Oklahoma. This year they lose OC Dana Holgorsen to West Virginia, and QB Brandon Weeden has been teaching the old offense to the new guy (what could possibly go wrong?). Losing Kendall Hunter is a big blow on top of that to the offense, which under Mike Gundy has been among the few in the Big 12 to keep a robust running game over the past five seasons. The road slate is not easy, though thanks to a quirk of the Big 12's switch to nine conference games, Bedlam will be held in Stillwater for the second straight year. This team is set up to take a step backwards on many fronts, and it could be too much to overcome.
Possible losses: 7 (Arizona, at Texas A&M, at Texas, at Missouri, Baylor, at Texas Tech, Oklahoma)
2. Texas A&M
The Aggies are the only preseason top 10 team that didn't win 10 games last year. As Bill Connelly pointed out to me yesterday, they also didn't even play like a true top 10 team during their six-game winning streak to end last year. They lit up bad defenses but struggled mightily against good ones, and overall, they didn't display a dominant streak I'd expect from a team with top 10 potential. This year it's a plus that they will play with Ryan Tannehill all season, but Von Miller is a huge loss on defense. This team that hasn't had high expectations in a long while, and they could easily start 2-2 with losses to Oklahoma State and Arkansas. If they somehow follow up those losses with another on the road at Texas Tech the following week, the bottom could fall out.
Potential losses: 7 (Oklahoma State, Arkansas, at Texas Tech, Baylor, Missouri, at Oklahoma, Texas)
I hate to tip my hand early, but I don't see a repeat of LSU's 11-2 record from a year ago. The Tigers were an amazing net +5 in close games, which is a credit to Les Miles with a hat tip to Derek Dooley. As I mentioned two days ago though, the defense has lost the best player at each level. Plus, losing a phenomenal return man and the kicker is not recommended for pulling out close wins. At this point I think Jordan Jefferson is what he is, which is an inconsistent quarterback. With potential rotation under center, question marks at receiver, and an incredibly unfortunate offensive coaching staff shakeup, I have a hard time seeing much improvement on offense. LSU was fortunate last year (luck: + 1.11), and in the toughest division of the toughest conference in football, it may be time to pay up for that this fall.
Potential losses: 6 (Oregon, at Mississippi State, at West Virginia, Florida, at Alabama, Arkansas)
A common thread through three of the four teams listed so far is coaching staff shakeups. The Cardinal lost its master builder Jim Harbaugh to the NFL, though to his credit he wasn't getting out of town just ahead of the cops like Pete Carroll was at USC. Will David Shaw be able to keep it going? I don't know. The main reason why Stanford isn't higher is because I'm not impressed by the Pac-12 in its opening season. If the conference slate was more difficult, this team certainly could be higher. I'm also on record for predicting a regression of Andrew Luck's numbers, so there's that too.
Potential losses: 5 (at Arizona, at USC, at Oregon State, Oregon, Notre Dame)
5. Florida State
I've already nitpicked this team to death, so hit up that article if you want to see 2011 FSU die by a thousand small cuts. Nitpicks were pretty much all I could come up with though. In addition, FSU was actually a 10-win team with bad luck last year (-1.53). So why aren't Jimbo's guys lower on the list? Largely it's because I don't think this team is as bulletproof as some others, and I think the ACC will be pretty good this year. Plus, I need to see a really excellent season from Mark Stoops before I can fully trust him. He and Mike never put together a dominant defense (or even top 25 or 30 defense) in Arizona, and while his FSU D last year was a big improvement, almost anything would have been over a defensive staff who the game had passed by. Of course, he's got much better players in Tallahassee than he ever did in Tucson.
Potential losses: 5 (Oklahoma, at Clemson, NC State, Miami (FL), Florida)
Overall I think Alabama's defense will prevent the team from losing more games than it did last year, but we've seen dominant defenses in the SEC go to waste before. Bama largely rates this highly on the list because of the trials and tribulations of being in the SEC West. The team doesn't have a clear quarterback coming in like it did with Greg McElroy in 2009, and if injuries bother Trent Richardson again, there's not much other than Eddie Lacy behind him. No one on the roster will approach Julio Jones' playmaking ability, and we're all going to find out how much Marquis Maze benefit from all the attention on Jones. If offense can't ever get things going, a lot of close losses could ensue.
Potential losses: 5 (at Penn State, Arkansas, at Florida, LSU, at Mississippi State)
I'm putting the Sooners this low against the better judgment of some of my favorite projection stats. They had good luck last year (+1.07), were net +5 in close games, and enjoyed a favorable turnover margin (+14). Plus, their mark of 3.68 in yards per point differential (YPP allowed minus YPP gained) was good but not great (24th in the nation). I keep them this low anyway because I think the defense will be better this year while the offense won't skip too much of a beat. Another huge factor: Bob Stoops is an ungodly 72-2 at home as a head coach. Oklahoma would basically have to go 1-5 in games away from Norman to finish unranked, a feat that it took an injured Sam Bradford to pull off in 2009. As long as Landry Jones remains upright, OU should finish in the polls.
Potential losses: 5 (at FSU, Missouri, Texas, Texas A&M, at Oklahoma State)
The Badgers would be higher on this list if not for the transfer of Russell Wilson, but they might not even be in the top 10 without him. I don't think he's a game changer in that he's the difference between eight and 12 wins, but he is a game changer in that you don't have to worry about quarterback anymore. Wisconsin always has a good offensive line, and it's absolutely stacked at running back this year. Wilson might only need to throw it 16-20 times per game to be a success. The defense sustained some losses, but it'll probably be good enough in what appears to be a down-ish Big Ten. This team might not be as dominant as it was last year, but it's a good bet not to drop more than three games even if Wilson tears an ACL.
Potential losses: 4 (Nebraska, at Michigan State, at Ohio State, Penn State)
9. Boise State
The Broncos play by a different set of rules than anyone else here, being from the MW
C. I think they can finish ranked with three losses if two are to Georgia and TCU while the third is a squeaker of an upset. Their reputation among coaches is that good right now. I don't think they can survive four losses, but the good news is that they almost certainly won't have to find out what that would feel like. The defense survived the loss of coordinator Justin Wilcox last year with no real drop off, and the offense will likely do fine without Bryan Harsin thanks to Chris Petersen still being around. Kellen Moore will suffer statistically for the same reason that Luck will on top of losing his best receivers, but he's good enough to keep everyone's favorite team in Idaho ranked all year.
Potential losses: 4 (Georgia, at Fresno State, Air Force, TCU)
A major caveat goes here: I offer no warranties on this placement if the NCAA turns up really bad evidence in regards to Will Lyles and blows up the program midseason. With that out of the way, I do think Oregon has some red flags despite being named least likely to finish the year unranked. It had only six starts missed to injury last year, and it had a favorable turnover margin (+13). However, its YPP differential (7.22, fifth nationally) was indicative a really good team, and it had a negligible luck factor (+0.31). The Ducks earned what they got last year, and despite some losses on the lines, they should be able to keep it up in the down Pac-12. Once the Ducks' offense gets rolling, and it eventually rolls against pretty much anyone in its conference, it becomes a landslide that covers over a multitude of sins elsewhere.
Potential losses: 3 (LSU, at Stanford, USC)