Compu-Picks 2011 Preview: Independents

Mr Pac Ten
Posted Jul 24, 2011

2011 Compu-Picks Previews Each 1-A League: Independents

Below is the preview for the Independents, consisting of two tables.

The first table outlines the projected rankings for each Independents team, sorted from best to worst, and then presents the following selected key stats:
Rank - Projected 2011 ranking, from 1 to 120
2010 Rank - 2010 ranking using the current compu-picks model
Prev 4 yr - ranking of the average rating from 2006-2009
Recruit Rank - ranking of past 4 years of recruiting (each year equally weighted), from
Recruit Trend - the difference between the past 3 years of recruiting and the previous 3, ranked from best to worst
Injuries - starts lost to injury during the 2010 season (from Phil Steele)
Turnovers - turnover margin during the 2010 season, from
Draft Losses - based on the 2011 draft
Ch - new head coach, per (1 indicates a new coach, . indicates no new coach)
Starters - returning offensive / defensive starters, per Phil Steele magazine (* if the QB returns), with some edits due to subsequent news

The second table shows the expected number of total wins for each team, as well as the odds for each potential number of wins, based on 2500 season simulation runs (note: a . indicates zero odds, while 0% indicates a non-zero probability that just rounds to 0%).

Projected ranking and key statistics

Team Rank 2010 Rank Prev 4 yr Recruit Rank Recruit Trend Injuries Turnovers Draft Losses Ch Starters
Notre Dame 10 22 50 11 107 30 1 9 . 9*/8
Brigham Young 14 61 15 42 15 43 1 0 . 10*/5
Navy 38 45 49 94 51 45 7 0 . 7/3
Army 88 75 112 96 75 21 16 0 . 6*/5

Projected Results - All Games

Team Rank SOS E(wins) 13 12 11 10 9 8 7 6 5 4 3 2 1 0
Notre Dame 10 16 9.38 . 15% 21% 21% 16% 10% 6% 4% 2% 2% 1% 1% 0% 0%
Brigham Young 14 61 9.22 . 10% 18% 21% 19% 14% 9% 5% 2% 1% 0% 0% . .
Navy 38 65 7.60 . 2% 5% 12% 18% 19% 16% 12% 7% 4% 3% 1% 1% 0%
Army 88 102 5.64 . 1% 2% 3% 7% 11% 13% 15% 14% 14% 10% 5% 4% 2%

Some commentary about the projections:

1) Compu-Picks is higher on the Irish than most (Athlon and Phil Steele had them at #6; the rest were lower than Compu-Picks), for a number of reasons: they recruit very well (not a surprise); they just lost only about half as much talent to the draft as usual; they have a very high number of starters returning; they suffered an unusually high number of injuries in 2010; and their 2010 turnover margin was very low compared to most of the other projected top teams this year. That said, their recruiting has been declining lately (though the 2011 class was a much-needed boost), and 2010 was already a notable outlier compared to where they had been. While further improvement is projected by Compu-Picks, it's far from certain.

2) Of course, the Notre Dame peg isn't going to be nearly as eyebrow-raising at the BYU peg. Unlike the Irish, almost no one is high on the Cougars (the only time I saw them in the top 25 was Wilner's January set, where he had them #13), even though unlike the Irish, BYU has actually been performing at a very high level for many of the past few years (though obviously not in 2010). Like Notre Dame, the Cougars have a number of positive indicators, such as: zero draft losses; a very high number of starts lost to injury in 2010; a low turnover margin in 2010; and a fairly high number of returning starters, and (unlike Notre Dame) recruiting that is rapidly improving (the 2011 class was poor, but the 2010 was the best in a long time and should pay dividends for a while). It's especially interesting that no one's high on them given how soft their 2011 schedule will be (Compu-Picks has it 61st, way worse than for most top teams), with only one projected top 15 opponent (TCU), and not many projected to be in the top 30. This team is absolutely capable of coming up with a 10+ win season.

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.

4) For three teams (Auburn, Oregon, UNC), you can see that they're projected to half a new coach. This was a manual adjustment I made to the data based on the off-field issues that each program is dealing with. A new coach is a negative predictive factor, so estimating a 50% chance of having a new coach makes an impact. For UNC, it's more that there's a pretty reasonable chance that they'll replace Butch Davis (or be forced to) before the season. For the other two, it's more a reflection of the possibility that the ongoing investigations will unearth more trouble, possibly leading to a coach replacement and/or current players being declared ineligible. For those two, estimating a 50% chance of a new coach is mainly a proxy for that possibility. It's admittedly arbitrary, but I believe that it's reasonable given the current climate.
I have also provided adjusted division (or league) odds in a number of instances. For the Pac-12 South, it shows the odds of each team winning adjusting for the fact that USC will be ineligible (the original calculation does not account for this). For various other instances, it would only be relevant if the team in question does in fact become ineligible for the division/league title. Should that not happen, you can ignore the adjusted odds.

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 Arkansas to be any better than LSU, 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. Please keep this in mind when looking at the distribution of win probabilities.

2011 Compu-Picks Blog

Questions, comments or suggestions? Email me at

Follow cfn_ms on Twitter