|TEAM||Adjusted Power Rating|
|San Diego State||68.00|
|San José State||62.40|
|Texas San Antonio||52.97|
|New Mexico State||43.18|
Cincinnati -20.5 vs UCF: I’ve got this game at 22 and will definitely take Cincy if there’s value in the line, so happy that I got it at -20.5. Cincy looks like they bounced back nicely against UConn, and I really have trouble seeing UCF put up much of a fight on the road.
Western Kentucky -21.5 at Old Dominion: I have this line higher at -22 to -26. I worry about WKU being banged up against LSU and having a letdown game on the road, but man ODU’s defense is so bad. They’re also 0-7 ATS, the only team without a a cover on the spread.
Tulsa -2 at SMU: I’ve got Tulsa in the -3 to -6 range in this game so will take the value at -2. SMU has struggled against great passing games in the 2nd half as we’ve seen their defense fall apart against Baylor, James Madison, ECU, and Houston. I expect them to start strong and fade late against Tulsa.
This Week: 5-2 (71%)
Season: 26-23 (53%)
Career: 85-62 (57.8%)
☆ Picks of the Week: 2-0
I managed to get off to a great start at 5-0, but the last two night games lost putting me at 5-2. Still a great week and I feel like I’m getting a better feel for a lot of these teams and seeing them improve/regress.
Toledo -9.5 at UMass (W): These guys had me on the ropes down 28-10 at halftime as Philip Ely threw 3 INTs in the first half. But they stormed back against a UMass defense that is known for being pretty terrible and covered the spread.
Duke +6 at Virginia Tech (W): It took you guys long enough, Duke! 4 OT later Duke pulls out the win at Virginia Tech – but Duke was clearly a great opponent and should not have been that big of an underdog.
Houston -19.5 at UCF (W): UCF put up a fight at first but really fell apart in the 4th quarter. Much similar to the Tulane game, Houston kept piling on the points even through the 4th quarter. Makes me wonder if he is trying to score style points in case the committee is looking at the them for a major bowl, or better. If that’s Tom Herman’s philosophy to run up the score, I plan to continue to bet on his team in these mismatched games.
Clemson -3 at Miami (W): So, they were a 3 point favorite, closed as a 7 point favorite…and won 58-0. Clemson looks very, very good. I will be surprised if Al Golden has a job on Monday.
North Carolina -12 vs Virginia (W): I kinda had some close games luck as UNC won by 13.
Texas A&M +6 at Ole Miss (L): I’m done betting on the Aggies for a while until they can figure some stuff out.
Florida State -6.5 at Georgia Tech (L): “What a time to be alive!” I got chills watching the end of that. Even though I lost money.
I was discussing before the season how this year was likely to be more chaotic due to more parity between the teams – not sure if I ever wrote anything substantial though, so I decided to revisit my simulation models from the beginning of the 2015 season, and compare to the simulation from the 2014 season.
Looking back to the beginning of the 2014 season, was one team “favored” (more than 50% odds) to win 11 games or more (Florida State). There were six other teams with a greater than 50% chance to win 10 games (Alabama, Oregon, Oklahoma, Wisconsin, Marshall, Bowling Green). The model was pretty correct as Florida State went 12-0,, and 3 of the 6 teams (Alabama, Oregon, Marshall) won 11 games, while Wisconsin hit 10 wins.
In 2015, the simulations predicted a much different scenario. Only one team was expected to win over 11 games, and only one other barely to break the 10 win mark. The end result: a lot of teams expected to go 9-3. Of course, it’s unlikely that each team actually goes 9-3 but the fact that all of those teams are clustered together showed that there was really no definitive group of “Football Overlords” this year that would have been comparable to the Alabamas, Oregons, and Florida States of 2014’s preseason.
|Year / Wins||2015||2014|
|12 Win Teams||0||0|
|11 Win Teams||1||1|
|10 Win Teams||1||6|
|9 Win Teams||13||15|
|8- Win Teams||113||106|
But some teams come out of nowhere!
But the problem is that the numbers above do not account for the inevitable teams that rise or fall. Looking back at last year, Oklahoma was the one favored to win 10+, yet TCU and Baylor ended up winning 11. These are very real scenarios that are already accounted for in the simulation. So I tracked and counted every team’s wins through about 20,000 simulations, labeling every team by conference, and aggregated them to count how many “playoff contender” teams would emerge at the end of the season.
Important Note: These numbers are as of the end of the regular season. There are many cases where 2 teams would finish 11-1 and both will be listed here, yet they would play each other in a conference championship. Still, the images give us a picture of what things might look like going into the postseason.
There are 3 types of teams that we will consider as “playoff contenders”:
- Undefeated Power Five Teams
- Undefeated Group of Five Teams
- One-Loss Power Five Teams
Note #2: There are many generalizations here – especially in the real of an undefeated G5 team being a playoff contender. If a team like Marshall last year went 12-0 it was widely agreed that they would not make it in, nor would they really be in consideration. However, many G5 teams at this point have resume-building wins that have established their legitimacy (Memphis over Ole Miss, Houston over Louisville, Temple over Penn State, Toledo over Arkansas) that an undefeated team of this sort should at least be considered. As a result though I have split the results into “Total” which is all inclusive, and “Total P5” which excludes G5 teams.
I have also added indices on the right to indicate the % of times the season ends in Low Chaos (0-4 teams), Medium Chaos (5-6 teams), or High Chaos (7+ teams). Last year for example was a Medium Chaos outcome (6 teams).
How accurate was the model’s prediction in 2014?
Pretty darn accurate. In 2014 the model predicted 1-2 undefeated P-5 teams , 0 undefeated G5 teams, and 4 one loss teams. The most likely outcome (23%) was a 6-team playoff race. In the end we had 1 undefeated and 5 one loss teams, for a total of 6 teams.
Ok, so what does this model predict for 2015 then?
In 2015 at the beginning of the season the model predicted 2 undefeated P-5 teams , 1 undefeated G5 teams, and 4-5 one loss teams. The most likely outcomes (19%) were either a 7-team or 8-playoff race.
Depending on how conference championships panned out, the final result of the simulation would of course show a lower number forecasted in the end – so who knows what will actually happen then. If that outcome happened, then many of the teams would be playing each other to determine who would win the conference, but that would not necessarily sort out into 4 teams. I personally am most intrigued by the prospect of a G5 team with a win over a major program (an SEC team maybe!) going undefeated and putting the committee into a very tough spot.
At the midseason mark, I wanted to go back and take a look at the teams that have surprised us the most – both for the better and for the worse. To do so I am comparing my forecasted power rating for 2015, compared to their current power rating in Week 8. I am not using raw improvement numbers because many of the teams that have improved the most like SMU and UConn are simply regressions to the mean – their teams were so terrible last year, that it was almost inevitable that they would improve in some regard.
As a result this is not, in my opinion, an accurate representation of how teams have outperformed our expectations. Sure, SMU improved, but we knew they would. Sure many teams have underwhelmed after losing their quarterback, but that was to be expected.
What we get are the results below, some of which definitely stand out:
Midseason Power Rating
|San Jose St||56.33||62.02||5.69|
|North Carolina State||72.97||77.68||4.71|
|San Diego State||61.79||63.73||1.94|
|New Mexico State||51.45||47.27||-4.18|
Toledo -9.5 at UMass (Foxboro): Another “home” game for UMass in Gillette Stadium that should feel more like a neutral site game. Even giving UMass 3 points for home field advantage, I have this game at -20.5? The line has since moved up to -14 but I would guess that it will climb higher.
Duke +6 at Virginia Tech: I have the line at like +1.5, give Duke credit for the bye and this is probably closer to pk. VA Tech has been a bit of a wild card especially with Brewer coming back into the picture, but overall I think Duke is definitely a more solid team.
Houston -19.5 at UCF: Fair line but I really like what Houston is doing. The improvement that Houston has made under Tom Herman is pretty evident, and I imagine they will continue to improve. UCF is a dumpster fire.
Clemson -3 at Miami: I have this game at -8. Current market line is -6.5 or -7 I think. I’m not too big on Miami, never really have been. And while they could upset Clemson, that’s not the most likely outcome here.
North Carolina -12 vs Virginia: Virginia is coming off a 3OT win. That means these guys must be extremely tired and beat up, which does not bode well for them going into the next week. Meanwhile UNC cruised by Wake Forest, a team that’s not that much worse than Virginia. I have the line at -15.5.
Texas A&M +6 at Ole Miss: You never know what you’ll get when you talk about a team coming off a tough loss (talking about Ole Miss, specifically). Do they come out and prove everyone wrong the next week? Is there discord in the locker room and they have terrible practices? Not really sure. But I do like A&M +6 because I have it at +2.5 and 6 seems like a lot of points. If I think about it a different way, would I take Texas A&M pk at home? Definitely.
Florida State -6.5 at Georgia Tech: Ok so this one doesn’t even agree with my handicaps. I have it set at -4 but intuitively it just…seems like FSU is so much better? I’m being punished for not following the numbers as the line moved against me to -5, but oh well.
This Week: 2-4 (33%)
Overall: 21-21 (50%)
☆ Picks of the Week: 2-0
☆ Western Kentucky -24 at UNT (W): Won this game pretty easily, but UNT actually managed to cover the -33 closing line thanks to a hailmary at the end of the first half and a late score in the last minute of the game.
Penn State +21 at Ohio State (L): Line closed at like 17 so had a good deal. I really though Penn State would be stout enough on defense to prevent Ohio State from scoring 38 points. Oh well.
Temple -17 at UCF (L): Four turnovers for Temple. They had a chance to cover late but chose not to. Line closed at -21 I think so still got a good deal, but Temple let UCF score too many points due to their turnovers.
Cincinnati +7 at BYU (L): This game looked so good early! I think Cincinnati was up 17-3 at one point. But Cincy totally blew this game, and capped it off by being down 14, with 4 minutes left, and all 3 timeouts, and choosing to punt and not go for it on 4th down to basically wave the white flag. Worst coaching decision I’ve seen in a long long time.
Louisville +10 at Florida State (L): Florida State just looked better. Hard to believe a rested Louisville defense gave up 41 points. Dalvin Cook and Everett Golson are a hard duo to stop though.
Oregon +3 at Washington (W): If Vernon Adams wasn’t playing this game could have gone in the opposite direction. But with him, I think Oregon can get itself back on track.
Oregon +3 at Washington: Vernon Adams is starting. That’s a game changer.
With 1:10 left on the clock in the 4th quarter, tied 45-45, TCU’s Trevone Boykin throws the ball deep to Josh Doctson. He catches the ball and runs towards the end zone, shaking off a final tackler at the 5 yard line. Doctson strolls in for the touchdown, giving the Frogs a 52-45 lead. Kansas State had no timeouts remaining at this point, and now gets the ball back and must drive down the field in an effort to tie the game.
If Kansas State is successful at leading a 1-minute march into the end zone (an outcome not that unlikely, as K-State had scored touchdowns on about of its possessions, albeit in much longer than 1 minute) then they would force the game into overtime, entering a situation where either team could win.
The game ends with K-State turning over the ball on their last possession, but I couldn’t help but wonder…was there a certain point that TCU already had a near 100% chance to win, that was passed up?
Let’s say that Doctson gets tackled at the 5 instead. TCU is now at 1st and goal, and K-State has zero timeouts. With 1:10 left, TCU can essentially run down the clock to 0:01, set up the ball in the middle of the field, and give Jaden Oberkrom essentially a PAT range kick that he has a 99% chance of making. And even if he missed, TCU would not lose, but merely go into OT.
So assuming the probability of each team winning in OT is fixed, then the only variable to compare is whether TCU is more likely to allow a touchdown in 1:10 of play, or miss a field goal. There are also ancilliary probabilities that K-State could score a TD quick to give TCU the ball back again, and that K-State could return a blocked extra kick for a TD.
Scenario 1: P(Make FG from 5 yard line) – P(Blocked and Returned Kick)
Scenario 2: P(Prevent K-State TD) – P(K-State Score Before End 4Q)
The probability of making from the 5 is about 0.57% based on historical data. While the average field goal kick has roughly a 2%-3% chance of being blocked, the majority of these are blocked because they are 30+ yard kicks that the kicker has to project the ball closer to a 45 degree angle to maximize distance. However from this range, the kicker can angle the ball higher to ensure it is not blocked.
While I have no data on what percentage of these are returned for touchdowns, let’s make a liberal guess of 25%. Based on these we get a total of:
Scenario 1: 99.64% TCU Win Odds
99.43% made field goal – TCU Wins
0.43% blocked FG no return – Overtime
0.14% blocked FG returned for TD – Kansas State Wins
|Team||Adjusted Power Rating|
|San Diego State||63.73|
|San José State||62.02|
|Texas San Antonio||55.55|
|New Mexico State||47.27|