Projected ranking and expected results
|Expected Wins||Projected League Results|
|Team||2012 Rank||2011 Rank||All Games||League Games||SOS||Div Finish||Division Odds|
Some notes and comments about the SEC and its teams:
1) Alabama - LSU is one of the more interesting questions of the preseason, with just about everyone having an opinion. As stated before (more than once), Compu-Picks takes Bama in this one. Bama loses more production (though after Mathieu's departure it's actually pretty close), loses much more to the draft, and had much better injury luck in 2011 than LSU. However, LSU had an almost certainly unsustainable turnover margin, isn't recruiting nearly as well as Bama (especially the most recent class), and simply hasn't been as good a program as Bama. To Compu-Picks at least, this really isn't a difficult choice.
And that's just in terms of power rating. Alabama also has a much easier SEC slate, drawing Mizzou and Tennessee from the East, compared to LSU's draw of Florida and South Carolina. That's a major difference, and is a big part of the projected gap in league records between the two.
2) With Arkansas projected to slip (largely due to Petrino being gone, though the fact that 2011 was already a material outlier compared to standard performance mattered as well), watch out for Auburn and, interestingly, Mississippi St as potential sleepers in the West. Auburn had a major down year in 2011, but has been a solid program, is recruiting fantastically well, and returns an enormous amount of talent. Mississippi St, meanwhile, returns much more production than their mere 12 starters would suggest, and unlike most teams near the top of the SEC, seems likely to improve their turnover margin rather than worsen. Also, Auburn has a reasonable East draw of Georgia and Mizzou, while Mississippi St has a fantastic East draw of Kentucky and Tennessee, arguably the easiest possible draw for a West team.
3) Compu-Picks isn't at all high on Texas A&M this year. While it's true that A&M had a lot of rough luck last year (turnovers, fumble luck, injuries, not to mention close game luck), they also have a new coach, 2011 was a positive outlier compared to previous history (not as much as 2010, but far better than 2007-2009), their recruiting trend is really bad (worst in SEC and bottom 15 in the country), and their offense is gutted, losing their starting quarterback, running back, and 2nd most productive receiver. Throw in a challenging schedule (not the nastiest in the SEC, but not very far off), and reaching six wins will be tough. Even worse, since they play two AA teams this year, they actually need to hit seven wins to make a bowl game. Compu-Picks is not optimistic about their chances of making that happen.
4) In the East, watch out for Florida. The Gators' recruiting slide seems to be over, with their numbers at least stabilizing as their amazing mid-decade recruiting run slips even further off the board. They also return an enormous amount of production on defense, though their offensive numbers show rebuilding potential. They also almost certainly will improve their turnover numbers, most likely by a substantial amount, which all by itself should herald meaningful improvement. Finally, while Urban Meyer is gone, the fact that this program has been absolutely oustanding for a while definitely comes into play here. A lot of the players from the 2008-2009 run are gone, but not all. There's more underlying talent here than people realize.
5) If it's not Florida in the East, it'll probably be Georgia. Drawing Auburn and Ole Miss is about as favorable an SEC slate as you can get, and this is a strong program with good recruiting numbers (though unlike many other SEC programs, they're actually slipping a bit rather than surging), a lot of returning production (though they do lose a bit to the draft), and overall a pretty solid resume.
6) South Carolina, meanwhile, looks iffier. An LSU-Arkansas draw isn't fun at all, but they also lost a lot of talent to the draft, and need to replace a lot of production on defense. Their recruiting numbers are solid but not spectacular, though they probably will get better injury luck in 2012. Overall, they're behind the projected top two, but not by so much that they couldn't end up the East's best team. However, thanks to the worst schedule draw of any projected East contenders, it's going to be a really tough road to actually make the title game even if, like last year, they really were the best team in the division. Don't be surprised if they come close, though.
7) As for the rest, Vandy and Mizzou look like the next best pair, with Mizzou projected to be a bit better quality-wise, while Vandy gets the projected record nod thanks to (like Georgia) drawing Auburn and Ole Miss instead of Mizzou's Bama and A&M. Vandy is intuitively more surprising, but their recruiting numbers have really surged, they return a lot of players, and they made nice strides last year despite the worst injury luck in the entire SEC. Mizzou, meanwhile, is basically projected to more or less hover around the usual spot quality-wise, with a worse record than usual just because the schedule is tougher than usual.
8) Compu-Picks didn't expect much out of Tennessee even before they lost Rogers. With him gone too, their numbers sink even lower. Don't be at all surprised if they miss a bowl for the second straight year.
The next two tables show key statistics and details underlying the projections, from prior history and performance to luck-related statistics to key indicators of incoming and outgoing talent. Below is a brief explanation of some of these items:
Rank - Projected 2012 ranking, from 1 to 124
2011 Rank - 2011 ranking using the current compu-picks model, from 1 to 120 (does NOT include the four 1-A newcomers)
Prev 4 yr - ranking of the average rating from 2007-2010, from 1 to 120 (does NOT include the four 1-A newcomers)
Injuries - starts lost to injury during the 2011 season, from Phil Steele
Fumble Luck - the number of net turnovers in 2011 due to fumble luck
Recruit Rank - ranking of past 4 years of recruiting (each year equally weighted), from scout.com
Recruit Trend - the difference between the past 3 years of recruiting and the previous 4, ranked from best to worst
Starters - returning offensive / defensive / special teams (kicker and punter) starters, per Phil Steele magazine (* if the QB returns), with some edits due to subsequent news
Returning Yards, Tackles, Int, Sacks, Lettermen - returning production and roster depth; lettermen taken from philsteele.com, with the other stats calculated from cfbstats.com.
Draft Losses - based on the 2012 draft
Key Statistics - Performance, Luck and Coaching
|Team||2012 Rank||2011 Rank||Prev 4 yr||Injuries||Turnovers||Fumble Luck||New Coach|
Talent Inflows and Outflows
|Team||Recruit Rank||Recruit Trend||Starters||Ret. Yards||Ret. Tackles||Ret. Int||Ret. Sacks||Ret. Lettermen||Draft Losses|
The next two tables show probability distributions for the projections, based on 5,001 season simulation runs. Please note that a . indicates zero odds, while 0% indicates a non-zero probability that just rounds to 0%. The first table breaks down results across all games, while the second breaks down results across league games only.
Projected Results - All Games
|Odds of Winning _ Games|
Projected Results - League Games
|Odds of Winning _ League Games|
There are a few important notes and caveats I need to make about this model:
1) Compu-Picks does not endorse implicitly or explicitly any form of illegal gambling. Compu-Picks is intended to be used for entertainment purposes only.
2) No guarantee or warranty is offered or implied by Compu-Picks for any information provided and/or predictions made.
3) This preseason model is primarily based on the main compu-picks model. Essentially, it attempts to predict how well a team will rate given its rating history, as well as a number of other data points, such as returning starters, draft talent lost, turnovers, recruiting, etc. This means, among other things, that the rankings are power rankings based on how good a team projects to be, as opposed to a more cynical (though accurate) model that attempts to project how the BCS will rank a team by making adjustments to favor those with easy schedules and punish those with tough schedules.
I have provided adjusted division (or league) odds in a couple of instances. For the Big Ten Leaders, it shows the odds of each team winning adjusting for the fact that Ohio St and Penn St will both be ineligible. The same is true for the ACC Coastal and North Carolina.
5) There is a substantial amount of noise in these projections, which is to be expected given the large number of unknowns (who will have good and bad luck with injuries, which young players will improve and which won't, how specific matchups will come into play, etc.). Right now the standard error is a bit over 0.2 on a scale of about -1 to +1. It's important to look at the projections with this in mind to get a sense of how material the projected differences are. Given a standard error around 0.2, it is safe to project Alabama to be a much better team than Mississippi St, but it is not safe to project Mississippi St to be any better than Arkansas, much less a lot better.
6) At this point, there are a number of model features that need to be investigated further. Chief among these is the distribution of extreme events. It appears that the model may be overstating the probabilities of extreme events, such as 12-0 or 0-12 records, or major underdogs winning their division/league. This overstatement has been reduced compared to last year's projections, but still likely exists to some degree. Please keep this in mind when looking at the distribution of win probabilities.
7) Since there is much less data available for the four 1-A newcomers, the power rating methodology has been more manual and arbitrary. As a consequence, I am somewhat less confident of the projections for those four teams than I am for the other 120 1-A members. Please keep this in mind when looking at the newcomers' projections.
Questions, comments or suggestions? Email me at firstname.lastname@example.org