Fiu, Cirminiello, Mitchell on TV - Campus Insiders | Buy College Football Tickets

Compu-Picks 2011 Analysis: Week Two

Mr Pac Ten
Posted Sep 11, 2011


Compu-Picks 2011 Analysis: What the system can learn from results of the first two weeks

This year, I published a number of Compu-Picks preview articles, which among other things included record projections for each team. These projections were based on 2,500 season simulations, and consisted of the average number of wins across all projections. But beyond just the overall record, the system also simulated every game played by every 1-A team in each of the 2,500 simulations. This means that, even though there isn't enough data to use Compu-Picks for weekly picks yet, I can still come up with some interesting data and results from the games so far.

With that in mind, I'll look at a couple of the more interesting I'll look at a couple of interesting teams so far. For each of these teams, I'll show their preseason projection, as well as what their projections would like look like given their record after week one and their record after week two. While everything in these tables is based on the projections run before any games have been played, you can see the impact of each win or loss on a team's total season projections. Let's start with two teams that have started on a hot 2-0 streak, Michigan and Texas.

In each of the two below tables, the original preseason projection is at the left, and each column to the right represents their projection given a win in week one (2nd column) and a followup win in week two (3rd column). For both teams, there wasn't much difference between the first and second columns. Each team was solidly expected to win their week one game, so the fact that they avoided a big upset prevented some of the major disaster scenario seasons. For the second week, though, each of them achieved a meaningful win (especially Michigan, who both Compu-Picks and the oddsmakers made an underdog against Notre Dame). And for each, being 2-0 increases their win projections nicely. Neither is projected as a true top team so far, as both still have win projections of under 9 wins, but things are certainly looking up from where they were at the start of the season.

I should emphasize that these tables are conditional expectations that don't reflect how each of their future opponents have done so far, or what the actual scores have been, just that there was a 1-0 or 2-0 start. So it is a simplification, though one that I still think is interesting to look at.

Michigan Win Odds and Expectations Preseason Start 1-0 Start 2-0
Total # Simulations 2500 2007 415
Win % vs
Western Michigan 80% 100% 100%
Notre Dame 19% 21% 100%
Eastern Michigan 97% 98% 100%
San Diego State 71% 75% 87%
Minnesota 75% 78% 91%
Northwestern 49% 53% 69%
Michigan State 21% 24% 35%
Purdue 64% 69% 84%
Iowa 52% 56% 69%
Illinois 48% 52% 67%
Nebraska 31% 35% 50%
Ohio State 17% 19% 33%
Total Expected Wins 6.24 6.78 8.84
Total Expected Wins After Week Two 5.25 5.57 6.84

Texas Win Odds and Expectations Preseason Start 1-0 Start 2-0
Total # Simulations 2500 2259 1136
Win % vs
Rice 90% 100% 100%
Brigham Young 48% 50% 100%
UCLA 38% 40% 49%
Iowa State 88% 90% 93%
Oklahoma 28% 30% 36%
Oklahoma State 44% 46% 54%
Baylor 74% 76% 82%
Kansas 85% 87% 91%
Texas Tech 63% 66% 73%
Missouri 36% 38% 44%
Kansas State 85% 86% 90%
Texas A&M 29% 31% 37%
Total Expected Wins 7.09 7.39 8.49
Total Expected Wins After Week Two 5.71 5.89 6.49

And now let's look at a couple of the 0-2 teams, Georgia and Notre Dame. Unlike the above pair of 2-0 teams, where the two teams' stories are very similar to each other, here the two tables tell a much different story from each other. For Georgia, the 0-2 start was not all that surprising; of the 2,500 season simulations, over 1,500 of them had the Bulldogs going 0-2. As a consequence, while the projection going forward is very ugly, it's not THAT much worse than it was to start with. Here it's more of a validation of Compu-Picks' original projections than a major surprise that forces a re-evaluation.

Notre Dame, on the other hand, was supposed to be a solid favorite in each of their first two games. Instead they're 0-2 and reeling. Of course, both losses were close, and the Irish seem like a team that could go on a run if they can stop beating themselves, but again, all this analysis sees is 0-2, not how those games went. And 0-2 tells an ugly story. If the Irish can rebound and beat Michigan St, the story starts to get a bit rosier, but right now that's a big if. Win or lose, I'm guessing that if I do this again next week the Irish will be back for more analysis.

Georgia Win Odds and Expectations Preseason Start 0-1 Start 0-2
Total # Simulations 2500 1810 1518
Win % vs
Boise State 28% 0% 0%
South Carolina 20% 16% 0%
Mississippi 42% 37% 34%
Mississippi State 51% 46% 43%
Tennessee 25% 21% 19%
Vanderbilt 68% 65% 62%
Florida 48% 44% 42%
New Mexico State 97% 96% 96%
Auburn 37% 33% 30%
Kentucky 62% 58% 55%
Georgia Tech 49% 44% 40%
AA 99% 98% 98%
Total Expected Wins 6.25 5.57 5.19
Total Expected Wins After Week Two 5.78 5.41 5.19

Notre Dame Win Odds and Expectations Preseason Start 0-1 Start 0-2
Total # Simulations 2500 353 122
Win % vs
South Florida 86% 0% 0%
Michigan 81% 65% 0%
Michigan State 78% 66% 51%
Pittsburgh 78% 61% 48%
Purdue 82% 69% 58%
Air Force 76% 59% 45%
Southern California 76% 63% 49%
Navy 86% 75% 65%
Wake Forest 87% 75% 61%
Maryland 83% 72% 60%
Boston College 87% 73% 58%
Stanford 38% 22% 12%
Total Expected Wins 9.38 7.00 5.07
Total Expected Wins After Week Two 7.71 6.35 5.07

And, finally, here are a couple 1-1 teams, BYU and TCU. For BYU, the win projection given a 1-1 start is very similar to the original projection, as both of the first two games were projected as being fairly close to tossups. Interestingly, given that they started 1-1, it was projected to be about a 2/3 chance that the win would have been Ole Miss about about a 1/3 chance that it would have been Texas. Given how close each of the two games was, as well as the fact that Ole Miss was the win and Texas the loss, that looks about right.

TCU is a bit of a different story. For TCU, each of the first two games were supposed to be pretty easy, especially Baylor, which was the game that they lost. Going forward, the projections aren't much different now than they were preseason, though that's mainly because most of TCU's games are against heavily overmatched opponents (the total record projection is less, since the 1-1 start was worse than expected). It's somewhat interesting to look at the Baylor and Air Force odds given a 1-1 start. Given a 1-1 start, the system thought that beating Air Force and losing to Baylor (what actually happened) was MUCH less likely than the other way around (this is also why 935 simulations had TCU start 1-1, but only 325 had them starting 0-1). Like I mentioned last week, Baylor's upset win was a big surprise for the model and caused a substantial re-evaluation of the Bears and their prospects. This is another reflection of that story.

Brigham Young Win Odds and Expectations Preseason Start 1-0 Start 1-1
Total # Simulations 2500 1619 1050
Win % vs
Mississippi 65% 100% 65%
Texas 52% 58% 35%
Utah 66% 71% 64%
Central Florida 88% 91% 89%
Utah State 97% 99% 98%
San Jose State 94% 97% 95%
Oregon State 62% 69% 62%
Texas Christian 31% 37% 30%
Idaho 97% 98% 98%
New Mexico State 99% 99% 99%
Hawaii 71% 76% 70%
AA 100% 100% 100%
Total Expected Wins 9.22 9.96 9.03
Total Expected Wins After Week Two 8.06 8.38 8.03

Texas Christian Win Odds and Expectations Preseason Start 0-1 Start 1-1
Total # Simulations 2500 325 935
Win % vs
Baylor 87% 0% 84%
Air Force 62% 47% 16%
Louisiana-Monroe 97% 91% 95%
Southern Methodist 85% 72% 79%
San Diego State 86% 74% 82%
New Mexico 99% 97% 99%
Brigham Young 69% 53% 62%
Wyoming 94% 87% 92%
Boise State 31% 19% 24%
Colorado State 97% 90% 97%
Nevada-Las Vegas 99% 98% 99%
AA 100% 99% 100%
Total Expected Wins 10.04 8.26 9.28
Total Expected Wins After Week Two 8.55 7.79 8.28

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.

2011 Compu-Picks Blog

Questions, comments or suggestions? Email me at cfn_ms@hotmail.com

Follow cfn_ms on Twitter

Related Stories
Legge's Thoughts: Dawgs Gift Wrap Win for USC
 -by DawgPost.com  Sep 11, 2011
Full Transcript: Coach Hoke Postgame
 -by GoBlueWolverine.com  Sep 11, 2011
Photo Gallery: ND vs. Michigan II
 -by GoBlueWolverine.com  Sep 11, 2011