College football upset spotting, week eight

• <50% (3-3 on the season (+0.36); expected record of 0.43-0.57; 43.3% chance of undefeated): #25 Memphis (at Houston) (43.3%)
• 50-59% (7-6 (-0.26); no games)
• 60-69% (15-7 (+0.58); 2.59-1.41; 17.65%): #24 LSU (at Ole Miss) (62.7%), #18 Michigan State (vs. Indiana) (63.3%), #13 Notre Dame (vs. #11 USC) (66.0%), #10 Oklahoma State (at Texas) (67.4%)
• 70-79% (19-4 (+1.71); 4.52-1.48; 18.26%): #23 West Virginia (at Baylor) (72.2%), #20 UCF (at Navy) (72.4%), #2 Penn State (vs. #19 Michigan) (76.2%), #17 USF (at Tulane) (76.6%), #15 Washington State (vs. Colorado) (77.3%), #9 Oklahoma (at Kansas State) (77.5%)
• 80-89% (23-3 (+0.95); 2.54-0.46; 60.70%): #21 Auburn (at Arkansas) (82.5%), #14 Virginia Tech (vs. North Carolina) (85.6%), #8 Miami (FL) (vs. Syracuse) (86.0%)
• 90-99% (51-2 (+0.19); 2.88-0.12; 88.39%): #5 Wisconsin (vs. Maryland) (91.0%), #1 Alabama (vs. Tennessee) (97.7%), #4 TCU (vs. Kansas) (99.3%)

Based on these numbers, we should expect about four upsets…however, if Memphis coming back to beat Houston means anything, we should shift our expectations to about 3.46, or right in the middle of three and four. I’m still going to roll with four upsets: Indiana over #18 Michigan State, Texas over #10 Oklahoma State, #19 Michigan over #2 Penn State, and Colorado over #15 Washington State.

NBA League Pass TV Guide (Week One)

I started this last season but didn’t have time during the season to update it. However, we’re going to run it back. Same idea as last year:

I and most other NBA fans live for Zach Lowe’s yearly League Pass Rankings. He misses on some and hits on others, but he’s largely right because he watches more of this sport than anyone else, and that includes scouts paid by teams to do so.

I toyed with the idea of doing this last year, but I figure this is the most quantifiable way to give recommended viewing games for all basketball junkies across the universe. Below is a spreadsheet with composite ratings based on both teams playing for each game in the first week of the NBA season. I’ve highlighted the highest-rated game or games for each day, along with a table on the front end showing Lowe’s ratings from best to worst. This should give you an idea of what games are the most watchable past opening night, and I’ll be doing this weekly throughout the season with tweaks to the rankings as need be. Those tweaks will be based on injuries, surprises, disappointments, and much more.

Here’s your first Weekly TV Guide for NBA League Pass of the 2017-18 season:

College football upset spotting, week seven

Yeah so this week isn’t very good but whatever it’s football.

• <50% (3-2 on the season (+0.83); expected record of 0.47-0.53; 46.8% chance of undefeated): #25 Navy (at Memphis) (46.8%)
• 50-59% (6-5 (-0.16); 1.10-0.90; 30.41%): #24 Texas Tech (at West Virginia) (54.2%), #21 Michigan State (at Minnesota) (56.1%)
• 60-69% (11-5 (+0.5); 3.92-2.08; 7.74%): #19 San Diego State (vs. Boise State) (60.6%), #6 TCU (at Kansas State) (61.9%), #10 Auburn (at LSU) (65.5%), #20 NC State (at Pittsburgh) (67.9%), #11 Miami (vs. Georgia Tech) (68.1%), #23 Stanford (vs. Oregon) (68.3%)
• 70-79% (17-3 (+1.9); 2.19-0.81; 38.82%): #17 Michigan (at Indiana) (71.6%), #13 Oklahoma (vs. Texas) (71.9%), #8 Washington State (at California) (75.4%)
• 80-89% (19-2 (+1.26); 4.31-0.69; 47.34%): #14 USC (vs. Utah) (80.7%), #15 Oklahoma State (vs. Baylor) (86.8%), #7 Wisconsin (vs. Purdue) (87.4%), #5 Washington (at Arizona State) (87.8%), #18 USF (vs. Cincinnati) (88.2%)
• 90-99% (47-1 (+0.87); 4.68-0.32; 71.8%): all remaining games

I’d warned about this for a couple weeks, but we finally got a couple of huge upsets that we’ve been waiting for: Michigan State over #7 Michigan (82.5%) and Iowa State over #3 Oklahoma (92.7%). We’re still running significantly behind expectations for the 70-99% upsets, though. 70-79% teams should be about 15-5 right now; that’s likely to regress to the mean at some point this season. Same for 80-89%. Because the teams in 90-99% are that much better than their opponents, I wouldn’t be totally surprised if we only got the one 90+% loss this year.

Anyway, as you can see, there isn’t a single Top 25 vs. Top 25 game this weekend. It’s a pretty weak slate, all told. However, considering the unusually light season for upsets so far – we’re running about 5.2 upsets below expectation just six weeks in, which is seriously astounding – we’re due for something stupid. I’m going out on a limb and projecting six Top 25 teams to lose this weekend: Memphis over #25 Navy, West Virginia over #24 Texas Tech, Boise State over #19 San Diego State, Kansas State over #6 TCU, Texas over #13 Oklahoma, and Utah over #14 USC. Why not! Let’s root for some mean regression this weekend.

College football upset spotting, week six

We already got one upset last night with NC State beating a Louisville team that’s very likely Lamar Jackson and 21 JAGs. However, that won’t be the only one this weekend:

This is a deeply unusual week – 1.5 seasons in, I’ve never had a week that gave three ranked teams as underdogs to unranked teams. It’s really weird, and I have a feeling we might have One Of Those Weeks. Your upset range is anywhere from three to seven within one standard deviation.

• Sub-50% (1-1; expected record 1.22-1.78): #20 Utah (vs. Stanford) (33.0%), #11 Washington State (at Oregon) (44.3%), #13 Miami (at Florida State) (44.8%)
• 50-59% (6-3 (+0.93); expected record 1.09-0.91): #17 Louisville (at #24 NC State) (53.8%), #21 Florida (vs. LSU) (55.0%)
• 60-69% (11-5 (+0.5); expected record 0-0)
• 70-79% (13-3 (+1.04); expected record 3.14-0.86): #22 Notre Dame (at North Carolina) (77.2%), #8 TCU (vs. #23 West Virginia) (78.4%), #19 San Diego State (at UNLV) (78.6%), #9 Wisconsin (at Nebraska) (79.8%)
• 80-89% (14-1 (+1.24); expected record 4.98-1.02): #4 Penn State (at Northwestern) (80.7%), #25 UCF (at Cincinnati) (81.2%), #7 Michigan (vs. Michigan State) (82.5%), #5 Georgia (at Vanderbilt) (82.6%), #16 Virginia Tech (at Boston College) (83.0%), #12 Auburn (vs. Ole Miss)
• 90-99% (42-0 (+1.44); expected record 5.57-0.43): Every other game

There’s already been one upset with #24 NC State over #17 Louisville. Along with those, give me Stanford over #20 Utah, Florida State over #13 Miami, UNLV over #19 San Diego State, and – why not – Northwestern over #4 Penn State. We’re due a shocker. If every group reverted to their mean this week, we’d see eight upsets. Chaos may be coming.

College football upset spotting, week five

Headed on an airplane somewhere…

• 50-59% (6-3 on the season (+0.93 above expectation): No games
• 60-69% (8-4 (+0.01): #5 USC (at #16 Washington State) (60.3%), #14 Miami (at Duke) (60.4%), #7 Georgia (at Tennessee) (60.7%), #13 Auburn (vs. Mississippi State) (69.2%)
• 70-79% (9-3 (+0.03): #2 Clemson (at #12 Virginia Tech) (70.5%), #15 Oklahoma State (at Texas Tech) (72.3%), #21 Florida (vs. Vanderbilt) (77.8%), #19 San Diego State (vs. Northern Illinois) (77.9%)
• 80-89% (11-1 (+0.83): #10 Wisconsin (vs. Northwestern) (84.3%), #4 Penn State (vs. Indiana) (86.6%), #18 South Florida (at East Carolina) (88.3%)
• 90-99% (37-0 (+1.23): Every other game

My picks are #16 Washington State over #5 USC, Duke over #14 Miami, and…I don’t know, Northern Illinois over #19 San Diego State. I suck at these this year so whatever. A lot of that is due to some insanely bad luck late in games (Kentucky-Florida, Iowa-Penn State, Tennessee-Georgia Tech, etc.), but…anyway, 2-12 on the year.

College football upset spotting, week four

As bad as I am at picking these games (1-9 now!), our system is working: it projected four Top 25 upsets last week. We got four exactly, and it was mostly my fault for picking the wrong games: the system correctly identified three potential areas for upsets with the 50-59% range (Memphis over #25 UCLA), 60-69% (Vanderbilt over #18 Kansas State), and 70-79% (San Diego State over #19 Stanford). It did expect a fourth upset from the 80-89% range, but we got that in 70-79%: Mississippi State destroying #12 LSU. So…we’re moving in the correct direction. That’s a positive! Amazingly, if I’d just picked every one of the 50-89% teams to lose last week, I’d have a better record (4-8) than I do on the season with my subjective picks. Fun times.

Anyway, here’s this week’s chart:

I mentioned previous weeks as being the potential “the best upsets happen on the worst weeks” weeks, but this might really be it. There’s two good games (the two Top 25 vs. Top 25 matches) and a whole lot of potentially close games that offer no serious appetite nationally. The ABC game (#4 Penn State at Iowa) got moved to 7:30 PM ET, which is ABC admitting they think this game’s pretty boring. Heck, the most interesting Top 25 vs. unranked opponent game of the week (#8 Michigan at Purdue) runs at 4 PM on FOX.

With all that said, this week has serious poll-toppling opportunities. Here’s the categories by win percentage for the week:

• 50-59% (5-2 (+1.02 above expectation) on the season, expected record of 1.08-0.92, 29.26% chance of 2-0): #17 Mississippi State (at #12 Georgia) (50.1%), #23 Utah (at Arizona) (58.4%)
• 60-69% (6-2 (+0.66) on the season, expected record of 2.67-1.33, 19.83% chance of 4-0): #22 San Diego State (at Air Force) (62.6%), #20 Florida (at Kentucky) (67.5%), #6 Oklahoma State (vs. #16 TCU) (67.6%), #24 Oregon (at Arizona State) (69.4%)
• 70-79% (5-3 (-0.99) on the season, expected record of 2.98-1.02, 30.84% chance of 4-0): #21 USF (vs. Temple) (71.6%), #7 Washington (at Colorado) (73.0%), #4 Penn State (at Iowa) (75.0%), #5 USC (at California) (78.6%)
• 80-89% (7-0 (+1.09) on the season, expected record of 4.26-0.74, 44.86% chance of 5-0): #14 Miami (FL) (vs. Toledo) (80.4%), #12 Florida State (vs. NC State) (82.8%), #8 Michigan (at Purdue) (86.5%), #15 Auburn (at Missouri) (87.5%), #3 Oklahoma (at Baylor) (88.7%)
• 90-99% (30-0 (+0.88) on the season, expected record of 6.65-0.35, 69.49% chance of 7-0): #25 LSU (vs. Syracuse) (90.0%), #1 Alabama (at Vanderbilt) (90.7%), #13 Virginia Tech (vs. Old Dominion) (93.8%), #18 Washington State (vs. Nevada) (95.0%), #2 Clemson (vs. Boston College) (97.5%), #19 Louisville (vs. Kent State) (98.9%), #10 Ohio State (vs. UNLV) (99.2%)

I mean…look at that. Even factoring in that some ranking systems probably have yet to fully factor in Missouri and Baylor’s suck, you can count…what, six potential upsets that are nearly impossible to envision? I can’t see Alabama losing to Vanderbilt, ever, and Clemson through Ohio State are safe, but…yeah, something weird’s gonna happen. We’re underperforming our usual upset totals, and regression to the mean is coming. It’s probably going to happen later in the year, but this week seems like it’s got a good shot.

Without rationalizing, here’s my picks, because I no longer care to rationalize picks that are 10% successful: #12 Georgia over #17 Mississippi State, Air Force over #22 San Diego State, Kentucky over #20 Florida, and Iowa over #4 Penn State. Why? Why not. Mostly because I can’t let myself pick Jim Harbaugh to lose to Purdue. I’d cry.

Look for four or five upsets this week, including a top five team potentially going down. The top five has just a 46.25% chance of staying intact for Week Five. Iowa over #4 Penn State? Sure! Why not. Cal beating #5 USC for the first time in like 14 years? It could happen! Vanderbilt? Nah, but they’re in a good place right now.

There will probably be an 80-99% range upset, too. Picking these games is a zero-sum game in itself because you look like an idiot when it’s wrong, but combined, they have just a 10.91-1.09 expected record this week and a 31.05% chance of going undefeated. There’s going to be a GIANT upset this week…let’s see who it is.

Two top ten teams losing this week? Absolutely. I have the top ten as a whole with a projected record of 7.57-1.43 and a chance of going 9-0 at just 19.59%. Typically, this means we get multiple big upsets, which…yes, let’s do it. Go Colorado!

What happens if it’s just a boring week with a couple upsets? Then we messed up somewhere along the line. The expected upsets for this week, within one standard deviation, are anywhere between 2.607 to 6.102. Two standard deviations (95% of outcomes): 0.859 to 7.849. We very rarely hit the second standard deviation, so that should put our range of upsets this week between 3 and 6.

East Tennessee high school football projections, week six

Last week’s record: 17-6, one short of our expected record of 18-5. On the year now, we’re 87-30 (.744), just one game under where I need to be to break the 75% threshold.

Thursday

Rockwood 32.1, Harriman 22.5 (Model 67.6%, S&P+ 71.1%)

Friday

Brentwood Academy 41.0, Knoxville Catholic 24.4 (Model 80.4%, S&P+ 83.2%)
Maryville 37.7, Fulton 23.4 (Model 76.1%, S&P+ 79.6%)
Dobyns-Bennett 29.3, Alcoa 27.8 (Model 52.8%, S&P+ 53.4%)
Farragut 39.3, West 9.9 (Model 99.9%, S&P+ 95.5%)
Greenback 30.6, Grace Christian 21.7 (Model 66.2%, S&P+ 69.6%)
South-Doyle 35.1, Carter 24.6 (Model 69.1%, S&P+ 72.8%)
Morristown East 27.2, Gibbs 25.8 (Model 52.6%, S&P+ 53.2%)
Oak Ridge 30.4, Sevier County 23.0 (Model 63.5%, S&P+ 66.6%)
Morristown West 38.1, Cherokee 22.1 (Model 79.3%, S&P+ 82.3%)
Bearden 47.8, CAK 10.9 (Model 99.9%, S&P+ 98.4%)
Central 32.8, Clinton 20.7 (Model 72.1%, S&P+ 75.9%)
Anderson County 37.9, Heritage 22.6 (Model 78.0%, S&P+ 81.1%)
Karns 29.7, Scott 29.1 (Model 51.1%, S&P+ 51.4%)
Austin-East 41.9, Halls 18.3 (Model 93.3%, S&P+ 91.4%)
Kingston 28.7, Stone Memorial 20.7 (Model 64.6%, S&P+ 67.9%)
Jefferson County 38.5, William Blount 13.0 (Model 96.7%, S&P+ 93.0%)
Trinity Academy 35.7, Gatlinburg-Pittman 22.5 (Model 74.2%, S&P+ 77.7%)
Powell 30.4, Lenoir City 22.1 (Model 65.3%, S&P+ 68.5%)
Loudon 32.8, Northview 23.3 (Model 67.4%, S&P+ 70.8%)
Cocke County 31.5, Pigeon Forge 29.7 (Model 53.3%, S&P+ 54.1%)

And this week’s games grouped by win probabilities:

• 50-59% (4 games, 1-3 last week, 18-13 on year): 2.12 expected wins-1.88 expected losses
• 60-69% (4 games, 3-1 last week, 9-10 on year): 2.73-1.27
• 70-79% (6 games, 2-2 last week, 17-4 on year): 4.48-1.52
• 80-89% (3 games, 5-0 last week, 22-2 on year): 2.47-0.53
• 90-99% (4 games, 6-0 last week, 22-1 on year): 3.78-0.22

 

College football upset spotting, week three

Eventually, we will hit on an upset. We’re 0 for 6 this year, though we did correctly nail that there would be an upset from the 60-69% range and at least one from 50-59%. It’s still one short of expectation. That typically regresses throughout a season. Either way, we’ll get our upsets in eventually.

• 50-59% (5-1 on the season, expected record of 0.53-0.47): #25 UCLA (at Memphis) (53.3%)
• 60-69% (3-1 on the season, expected record of 2.62-1.38): #18 Kansas State (at Vanderbilt) (62.7%), #24 Florida (vs. #23 Tennessee) (63.5%), #3 Clemson (at #14 Louisville) (67.5%), #9 Oklahoma State (at Pittsburgh) (68.0%)
• 70-79% (3-1 on the season, expected record of 3.00-1.00): #19 Stanford (at San Diego State) (71.2%), #12 LSU (at Mississippi State) (72.3%), #10 Wisconsin (at BYU) (78.1%), #4 USC (vs. Texas) (78.6%)
• 80-89% (4-0 on the season, expected record of 2.54-0.46): #22 USF (vs. Illinois) (80.0%), #21 Washington State (vs. Oregon State) (85.8%), #20 TCU (vs. SMU) (88.3%)
• 90-99% (21-0 on the season, expected record of 8.70-0.30): every other game

Considering I’m 0 for 6 on the season so far, I might as well shoot for 0-7 or worse. Memphis defeating UCLA is a fairly realistic proposition (two of our eight models say it’ll happen and seven of eight give Memphis a 40% shot or greater), and I’ve no clue why UCLA is ranked other than people politely forgetting how absolutely, utterly awful they looked for 2.5 quarters against Texas A&M. Plus, if I screw this pick up, they play at noon so it can get over with early.

We need three more upsets, and, because I’m a hopeful person, we’re picking one from each group. On average, we’re looking at about 1.3 upsets from the top ten teams and 2.3 from 11-25. One of them is Memphis over UCLA; the other two are #14 Louisville over #3 Clemson and…wait, BYU over #10 Wisconsin? Uh…sure!

I’m rationalizing this with the following: the top 12 teams playing this week (FSU’s off) have an expected record of 10.45-1.55. The remaining nine favorites are expected to go about 6.95-2.05. If we’re to follow this, we need two top 12 upsets. The reigning Heisman Trophy winner gets a Clemson team at home that’s yet to force a turnover in front of a bonkers crowd. Good enough for me.

The other one’s borderline impossible to rationalize because BYU has looked like absolute butt thus far, but Wisconsin (minus Jonathan Taylor Tailback) has yet to look consistently good on offense thus far. This could just as easily be a Wisconsin blowout as it could be a weird, wonky 17-14 BYU win. It’s worth the risk.

Forcing in two top 12 upsets leads me to leave out the San Diego State over #19 Stanford pick I’d hoped to use. We need a new one, and, wildly enough, we’re going to go for an 80-89% game. The 80-99% games this week have an expected record of 11.25-0.75 and, combined, have about a 44.3% chance of going 12-0. That’s still pretty good, but it’s more likely than not we get a serious upset loss. It’s also significantly more likely it’s an 80-89% team.

I only seriously considered two games from this bracket: Oregon State-#21 Washington State and Illinois-#22 USF. We’re rolling with Illinois over #22 USF and I’m not confident. Oregon State really should be 0-3, because FCS Portland State completely outplayed them in every sense of the word. Illinois just beat what I thought could be a 12-0 Western Kentucky team and it wasn’t a fluke. By recruiting rankings, Illinois has more talent. By coaching rankings, Charlie Strong was one of the very worst by CFB Matrix Coach Effect in the last decade. Go Illini!

I very rarely pick a 90-99% upset, but if there’s going to be one it would have to be an absolute shocker: something like East Carolina over #16 Virginia Tech or Air Force beating #7 Michigan. Neither is going to happen in all likelihood, but both did have at least one model peg them as less than 90% favorites. Also interesting was the idea of Samford over #13 Georgia. Sagarin Ratings had that game at 83.5%, which is crazy, but also lends respect to a Samford team expected to be a strong contender in the Southern Conference.

Good luck on your picks this week, and wait for the first national college football writer to talk about how the worst weeks have the best games. Or something like that.

East Tennessee high school football projections, week five

Finally, some correction in the projection world. After a bizarre first three weeks that had teams in the 50-59% range going 15-8 but those in 60-69% going 2-9, these started to regress to a theoretical mean: 50-59% teams went 2-2 and 60-69% teams went 4-0. In fact, our picks went 21-2 last week, bring us to 70-24 (74.5%) on the year. That probably won’t happen again, but it shows we’re heading down the right path. The goal by year’s end should be for this projection system to hit on about 75% of games. We’re right there!

Here’s this week’s picks:

Thursday

Sevier County 33.3, South-Doyle 19.0 (Model 76.1%, S&P+ 79.5%)

Friday

Webb 32.2, Grace Christian 17.6 (Model 76.6%, S&P+ 80.1%)
Alcoa 45.5, Pigeon Forge 6.7 (Model 99.9%, S&P+ 98.8%)
Morristown West 24.6, Hardin Valley 21.9 (Model 54.9%, S&P+ 56.1%)
Science Hill 35.0, Bearden 19.4 (Model 78.5%, S&P+ 81.7%)
Farragut 40.4, Jefferson County 16.8 (Model 93.2%, S&P+ 91.3%)
Campbell County 37.2, Gibbs 16.3 (Model 88.2%, S&P+ 88.7%)
Knoxville Catholic 47.4, Soddy-Daisy 16.6 (Model 99.9%, S&P+ 96.3%)
Oak Ridge 25.6, West 19.6 (Model 61.1%, S&P+ 63.9%)
Ooltewah 45.5, William Blount 13.2 (Model 99.9%, S&P+ 97.0%)
Fulton 48.0, Karns 11.5 (Model 99.9%, S&P+ 98.3%)
Central 35.6, Seymour 15.9 (Model 86.1%, S&P+ 87.2%)
McMinn County 29.2, Heritage 21.3 (Model 64.4%, S&P+ 67.5%)
Rhea County 35.1, Lenoir City 16.0 (Model 85.0%, S&P+ 86.7%)
Chattanooga Christian 29.4, CAK 25.9 (Model 56.5%, S&P+ 58.3%)
Gatlinburg-Pittman 35.4, Scott 24.8 (Model 69.4%, S&P+ 72.9%)
Carter 37.4, Halls 26.0 (Model 70.9%, S&P+ 74.5%)
Morristown East 28.5, Tennessee 25.0 (Model 56.5%, S&P+ 58.3%)
Anderson County 46.7, Howard 8.0 (Model 99.9%, S&P+ 98.7%)
Kingston 33.6, Northview 25.2 (Model 65.4%, S&P+ 68.7%)
Mt. Juliet Christian 30.9, King’s Academy 29.1 (Model 53.4%, S&P+ 54.5%)
Powell 32.6, Clinton 26.5 (Model 61.3%, S&P+ 63.9%)
Loudon 29.2, Signal Mountain 15.0 (Model 76.0%, S&P+ 79.4%)

And S&P+ win probabilities:

• 50-59% (4 games, 2-2 last week, 17-10 on year): 2.21 expected wins-1.79 expected losses
• 60-69% (4 games, 4-0 last week, 6-9 on year): 2.64-1.36
• 70-79% (4 games, 5-0 last week, 15-2 on year): 3.06-0.94
• 80-89% (5 games, 8-0 last week, 17-2 on year): 4.24-0.76
• 90-99% (6 games, 2-0 last week, 16-1 on year): 5.80-0.20

The teams from 70-99% have an expected record on the year of about 45-8; they’re three games ahead of that now at 48-5. The 70-79% group is especially overperforming at 15-2: their expected record is about 13-4. I’d guess that you see at least one significant upset this week and probably a few more smaller ones. The expected record of favorites this week is 18-5.

College football upset spotting, week two

First off, an apology: our normally pretty good system went a solid 0-3. It required Tennessee to get outgained by nearly 300 yards and win, along with West Virginia failing to convert in the red zone, but it still went 0-3. This week, we’ll get at least one upset correct. Maybe.

• 50-59% (2-0 on the season, expected record of 2.32-1.68): #23 TCU (at Arkansas) (54.4%), #6 USC (vs. #14 Stanford) (57.6%), #17 Louisville (at North Carolina) (59.6%), #24 Notre Dame (vs. #15 Georgia) (59.9%)
• 60-69% (2-0 on the season, 1.38-0.62): #2 Ohio State (vs. #5 Oklahoma) (67.9%), #3 Clemson (vs. #13 Auburn) (69.6%)
• 70-79% (2-1 on the season, 0.71-0.29): #20 Washington State (vs. Boise State) (70.8%)
• 80-89% (3-0 on the season, 0.81-0.19): #4 Penn State (vs. Pittsburgh) (81.0%)
• 90-99% (11-0 on the season, 9.80-0.20): every other game

Yet again, no 90+% game is worth discussing. There literally isn’t a single one of these teams with a model that projects them to win by less than 25 points. Not interesting at all! Conference play will even this out. Anyway, two key notes: the 50-69% grouping last week was expected to go 2.47-1.53; they went 4-0. A 70-79% team lost, but it was the last one I would have expected. The 50-69% teams will not go undefeated this week; I can promise this. There’s a 5.3% chance they go 6-0.

Alright, let’s pick some losers. The most surprising part of our work this week is the following: the models give the top four teams (#5 Oklahoma left out as they are playing #2 Ohio State) just a 38.1% chance of going 4-0 with 0.82 expected losses. I highly doubt #1 Alabama will drop to Fresno State, and while it did happen last year, Pittsburgh beating a top ten Penn State team two years in a row seems not right. So it’s between two 60-69% games: #2 Ohio State vs. #5 Oklahoma and #3 Clemson vs. #13 Auburn.

These are basically a coin flip, and I’m ready to lose on another toss-up, so I’m taking #13 Auburn over #3 Clemson. Why? Well, consider it two-fold: Auburn’s got a recruiting talent advantage over Clemson, and when removing the largest outlier model for both this game and OU-OSU, this is a slightly more likely event (66.9% vs. 69.9%). There’s a 53% chance that either #2 or #3 lose on Saturday night; good enough for me.

We still need two more upsets, and both are likely from the 50-59% range. We need one more from the Top 25 vs. Top 25 grouping. How about #14 Stanford over #6 USC? This knocks out our necessity of finding an upset in the Top 15. Again, after removing the outlier model, this one was a slightly more likely event than Georgia defeating Notre Dame. I also like this one because it hits on some concerns with USC: the significant schematic advantage Stanford’s offensive staff likely has over USC’s defensive staff, plus USC’s significant struggles with a Group of Five team in the first game. Perhaps USC’s defense ends up being pretty good, but I’m not yet sold.

For our final upset, we need one from the low end of the Top 25. It needs to be an unranked team. We’ve got two contenders: #17 Louisville losing at a mediocre North Carolina team or #23 TCU losing at a probably good-ish Arkansas team. You can probably guess which way I’m swinging here: Arkansas over #23 TCU. Arkansas, based on stats from CFB Matrix, has a large coaching advantage on offense versus defense over TCU. Plus, it’s a home game, the first CBS game of the year, and Arkansas has both a small recruiting talent advantage and doesn’t have to play the reigning Heisman Trophy winner.

To recap, our picks this week are #13 Auburn over #3 Clemson, #14 Stanford over #6 USC, and Arkansas over #23 TCU. Other games considered: #15 Georgia over #24 Notre Dame, North Carolina over #17 Louisville, #5 Oklahoma over #2 Ohio State, and Pittsburgh over #3 Penn State. Three upsets should come from this group of eight games. If there is a fourth upset, look for Boise State over #20 Washington State.