The wait is almost over. It has been almost 8 full months since the Spartans last took to the field, but on Friday night, that will all come to an end. Over the last few years, been posted a weekly preview and summary of the college football action, laced with just a little bit of math, and this year I figured I would give it another go.
First up for the Spartans on Friday night are the Aggies of Utah State. Based on the preseason rankings, USU has a consensus ranking of #78, making this, in theory, MSU's 3rd softest opponent on the schedule. The Aggies are right behind Maryland (68) and just ahead of Rutgers (86) and CMU (112) and they do have a history of given Power 5 teams trouble. Just ask Wisconsin. Utah State hails from the Mountain West Conference, where they are projected to be in a dog-fight for 2nd place with Wyoming (80) in the Mountain Division, with Boise State (21) being the overwhelming favorite (my math says that Boise has a 96% chance to win the division with both USU and WY having only a ~2% chance each). Pretty much the only other thing I know about the Aggies is that every time I think of their school's name, I think of this bad, obscure techno song from the 90s. No really. It's a curse.
I think most of us would agree that this week is mostly about MSU. With so many players back this year and with expectations much higher than last year, it will be interesting to see what the 2018 version of the Green and White will look like. My sources report that the Vegas line opened at MSU -25, which suggests a 97% chance that MSU wins, based on historical data. Are you 97% sure that MSU doesn't screw this one up somehow? Yeah, me neither. I certainly expect MSU to win, but I have a feeling that there will be some tense moments and maybe some grumpy fans by the end of the evening who did not get the blowout that there were expecting. That all said, my spreadsheet still likes MSU to cover (-30.3).
National Overview
In this space, it is now my tradition to present a few different pieces of data each week that I derive from my spreadsheet algorithm. My spreadsheet allows me to project point spreads for all games involving two FBS-level teams. Early in the season, I anchor each team's power ranking to one derived from a consensus of preseason rankings until 4-5 weeks into the season when the teams are "connected" enough for me to cut the preseason numbers loose. For Week 1, essentially all the projected spreads are based purely on the preseason rankings.
With this information, I typically make several predictions. These include some specific upsets, a total number of upsets for for the week, and teams that I expect to cover or not cover, depending on how far my algorithm's prediction differs from Vegas. After each week, will also compare my performance against the spread (ATS) and my upset predicting power to ESPN's FPI metric, since they like to talk about it so much. I will present the data for all games graphically (see below) and I will summarize the upset picks in a table.
Just for reference, my target performance for these predictions is simply to be over 50% for picks against the spread and over 40% for upset picks. Last year, I was 50.3% ATS overall, 52.8% for my "highlighted covers" and a surprising 48.1% for upsets. In 2017, I have ESPN's FPI at 53.1% ATS and 50.0% for upsets. But, I will add that my metric made twice as many upset picks as the very conservative FPI. Finally, last year I tried fun experiment where I also used a statistically weighted random number generator (nick-named "Rand()y") to also pick upsets each week. Poor Rand()y only got 36.2% of his upset picks correct and decided to retire.
Moving on the Week 1, the chart above shows my picks. In total, I like 8 teams to cover (Ohio State, Wisconsin, South Carolina, Oregon, USC, Boise, and Colorado) and oddly, one team not to: Alabama. In other words, my spreadsheet decided to open up the season by picking a fight with Saban. Great.
As for upsets, I have 5 total, which as summarized both in the figure above and the table below. The most notable of which are Northwestern over Purdue, Texas Tech over Ole Miss, and Notre Dame over Michigan. More on that contest later. As for the FPI, it took 4 of the same 5 upsets that my algorithm likes, but it favors Washington over Auburn instead of Wyoming over Washington State. It should be noted, of course, that none of these games have huge spreads, so they are all essentially toss-ups in the first place. That said, I also project that in total, there will be 9.2 ± 2.5 upsets in Week 1.
Similar to the Big Ten, essentially half of the SEC decided to partake of some sweet, sweet cupcake action this weekend, except they prefer their carb intake to take the form of FCS-level teams. Arkansas, Florida, Georgia, Mississippi State, Missouri, and Texas A&M all fall into the category, and South Carolina (-29.5) scheduled Coastal Carolina, a Sun Belt team who is so new to the FBS, the logos on their field are likely still damp. (It's the humidity). Also, Central Michigan is not likely to give Kentucky (-17) much trouble. But, the other six games are a bit more intriguing.
SEC doormat Vanderbilt (-6) is hosting Middle Tennessee State, and my math likes the Blue Raiders to cover, if not win. Trust me, Vandy, getting upset by Middle Tennessee is NOT FUN. But, it is the other 5 games that are the real must-see-TV this weekend as they all feature an SEC team playing another Power 5 team on a neutral (more-or-less) site. In Houston, Texas Tech will square off against Ole Miss (-1). Neither team is picked to be competitive in their respective leagues, but this game looks to be competitive, and my algorithm likes the Red Raiders in an upset. Charlotte will play host to a match between West Virginia (-8) and Tennessee, which my data suggests will not be as close as Vegas thinks. In Arlington, Miami (-3) will tangle with LSU in JerryWorld. If LSU is able to upset the Canes, it would be a major blow to a team that has a surprisingly reasonable shot at the playoffs. The defending champs, Alabama (-25.5) will travel to Orlando to face Louisville. Although Vegas thinks this will be a blowout, my spreadsheet predicts the Cardinal will cover easily. Not win, mind you, but just keep it under 20.
But the main event has to be a match-up of Washington vs. Auburn (-2.5) in Atlanta. With both teams entering the season in the Top 10 and both contenders in their leagues, this will be the first real heavy weight battle of the year. My spreadsheet also likes the Tigers, but only by the slimmest of margins (0.6 points).
ACC
Much like the Big Ten and SEC, the ACC has a handful of laughers on the schedule, sprinkled in with some interesting contests. Clemson, Georgia Tech, NC State, Pitt, and Virginia all open with FCS opponents, while Boston College (-18) and Wake Forest (-8) are double digit favorites against teams in the Bottom 40 of the FBS. Duke is -11 as a host to Army, and I mentioned above Louisville (+25.5) and Miami's (-3) big tests against Bama and LSU.
The three remaining contests feature two games that should be entertaining, if not impactful, and one that is clearly the biggest conference game of the weekend. In the non-impactful category, we have North Carolina going out to Berkeley to face the Cal Bears (-5) and Syracuse (-6) traveling to Kalamazoo to face Western Michigan. My spreadsheet does not pick an upset in either contest, but it likes WMU to cover (+2.4), and I am going to issue an upset watch for the Orange, as conditions seems favorable for the formation of an upset.
But, one of the biggest games of the weekend is the ACC opener featuring Virginia Tech at Florida State (-5.5). The Seminoles started the season last year as a consensus Top 5 team, yet injuries piled up early and they limped to a 6-6 regular season record. Virginia Tech is coming off a 9-3 regular season and some of my calculations suggest that the Hokies were better last year than we thought and might just be the surprise team of the ACC this year. My calculations also suggest that if the Hokie score the upset, they will have slightly better odds than Miami (49.5% to 46.4%) to win the ACC Coastal Division. However, a Florida State win could restore their confidence. Considering Clemson needs to travel to Tallahassee this year, one more win might be all it would take to shift the balance of the Atlantic Division as well.
Big 12
The Big 12 took a look at some of the scheduling in the other conferences this week and thought, "hold my beer." A full 60% of the 10-team Big 12 are facing FCS opponents, and 3 of the remaining 4 teams somehow managed to schedule Power 5 teams outside of the Top 50 in the previously mentioned Texas (-8.5) at Maryland, Texas Tech (+1) vs, Ole Miss and West Virginia (-8) vs. Tennessee contests. So, ironically, the highest ranked opponent a Big 12 team will face is C-USA favorite Florida Atlantic (preseason rank #36) who is travel to Norman to face the Sooners (-21). My math says that spread is a bit high (I peg it at -15.7), and my gut says that if you want to pick a crazy upset for Week 1... look here.
Pac 12
Due to the West Coasters generally dismal Bowl performance last year, I have demoted the Pac 12 to the last slot in my weekly analysis (Congrats Big 12!). As their first act of penance, they only put one FCS team on the schedule this week. That honor goes to Utah, who actually did win their Bowl Game last year, so that all seems fair. Honestly, there is not a whole lot to get excited about, as the Arizona schools, Oregon, Stanford, UCLA, and USC are all better than 2 TD favorites (although that San Diego State at Stanford (-15.5) game is a bit intriguing...), and Oregon State is close to a 40-point dog to the Buckeyes. The aforementioned UNC at Cal (-5) game should be watchable, and I suppose the Colorado (-6) vs, Colorado State game is always fun. I even have a predicted upset to watch for when Wazzou travels to Wyoming (+3.5).
But, when it comes down to it, the biggest game of the week is the previously mentioned match-up between Washington and Auburn. This game seems much bigger for Washington, honestly, as a loss would not just hurt their playoff chances, but it could effectively sink the whole Pac 12. Outside of Stanford or USC beating Notre Dame or USC beating Texas later in the season, there just aren't that many chances for the Pac 12 to gain a signature win. If Notre Dame were to also lose to Michigan, the entire conferences hopes of a quality victory would rest on the idea that Texas doesn't also suck. Good luck with that. Considering the conference went only 1-8 during Bowl Season, this is a game the Huskies really need to win, both for themselves, and for their conference.
Independents and Group of Five
We already talked about the Fighting Irish, so what about the Group of Five teams that could make a run at the New Years Six Bowls? In most years, Week 1 offers an interesting test for at least one of those schools. Well... not this year. I did mention the Florida Atlantic (+21) at Oklahoma game previously and other than that, unless you are really interested in Boise State (-10.5) vs. Troy, I suggest we just wait until next week.
That is all I have for now, so as always, Go State, Beat the Aggies!
First up for the Spartans on Friday night are the Aggies of Utah State. Based on the preseason rankings, USU has a consensus ranking of #78, making this, in theory, MSU's 3rd softest opponent on the schedule. The Aggies are right behind Maryland (68) and just ahead of Rutgers (86) and CMU (112) and they do have a history of given Power 5 teams trouble. Just ask Wisconsin. Utah State hails from the Mountain West Conference, where they are projected to be in a dog-fight for 2nd place with Wyoming (80) in the Mountain Division, with Boise State (21) being the overwhelming favorite (my math says that Boise has a 96% chance to win the division with both USU and WY having only a ~2% chance each). Pretty much the only other thing I know about the Aggies is that every time I think of their school's name, I think of this bad, obscure techno song from the 90s. No really. It's a curse.
I think most of us would agree that this week is mostly about MSU. With so many players back this year and with expectations much higher than last year, it will be interesting to see what the 2018 version of the Green and White will look like. My sources report that the Vegas line opened at MSU -25, which suggests a 97% chance that MSU wins, based on historical data. Are you 97% sure that MSU doesn't screw this one up somehow? Yeah, me neither. I certainly expect MSU to win, but I have a feeling that there will be some tense moments and maybe some grumpy fans by the end of the evening who did not get the blowout that there were expecting. That all said, my spreadsheet still likes MSU to cover (-30.3).
National Overview
In this space, it is now my tradition to present a few different pieces of data each week that I derive from my spreadsheet algorithm. My spreadsheet allows me to project point spreads for all games involving two FBS-level teams. Early in the season, I anchor each team's power ranking to one derived from a consensus of preseason rankings until 4-5 weeks into the season when the teams are "connected" enough for me to cut the preseason numbers loose. For Week 1, essentially all the projected spreads are based purely on the preseason rankings.
With this information, I typically make several predictions. These include some specific upsets, a total number of upsets for for the week, and teams that I expect to cover or not cover, depending on how far my algorithm's prediction differs from Vegas. After each week, will also compare my performance against the spread (ATS) and my upset predicting power to ESPN's FPI metric, since they like to talk about it so much. I will present the data for all games graphically (see below) and I will summarize the upset picks in a table.
Just for reference, my target performance for these predictions is simply to be over 50% for picks against the spread and over 40% for upset picks. Last year, I was 50.3% ATS overall, 52.8% for my "highlighted covers" and a surprising 48.1% for upsets. In 2017, I have ESPN's FPI at 53.1% ATS and 50.0% for upsets. But, I will add that my metric made twice as many upset picks as the very conservative FPI. Finally, last year I tried fun experiment where I also used a statistically weighted random number generator (nick-named "Rand()y") to also pick upsets each week. Poor Rand()y only got 36.2% of his upset picks correct and decided to retire.
Moving on the Week 1, the chart above shows my picks. In total, I like 8 teams to cover (Ohio State, Wisconsin, South Carolina, Oregon, USC, Boise, and Colorado) and oddly, one team not to: Alabama. In other words, my spreadsheet decided to open up the season by picking a fight with Saban. Great.
As for upsets, I have 5 total, which as summarized both in the figure above and the table below. The most notable of which are Northwestern over Purdue, Texas Tech over Ole Miss, and Notre Dame over Michigan. More on that contest later. As for the FPI, it took 4 of the same 5 upsets that my algorithm likes, but it favors Washington over Auburn instead of Wyoming over Washington State. It should be noted, of course, that none of these games have huge spreads, so they are all essentially toss-ups in the first place. That said, I also project that in total, there will be 9.2 ± 2.5 upsets in Week 1.
Big Ten
All 14 Big Ten teams are in action this weekend, but all but four of the teams are involved in games with spreads in the double digits, including MSU. Ohio State (-37.5), Wisconsin (-32.5), Penn State (-23.5), Nebraska (-20.5), Minnesota (-17.5), and even Rutgers (-18.5) and Illinois (-17) are all big favorites at home, likely because only Penn State's opponent (Appalachian State, #75) was in the Top 100 of the consensus preseason rankings. Strong cupcake scheduling, Big Ten brethren. Oddly, my spreadsheet likes all 7 teams to cover, as it tends to really love the favorite, especially early on.
In the middle tier of games, I have a few that I will keep my eye on. Indiana (-10) starts the season on the road at Florida International. This spread is a little tighter that my model suggests (-16) which gives me pause, but at the same time, IU is one of the teams I thought might be an over-achiever this year ("hidden treasure" perhaps) based on my post-season analysis of last year's data. I am curious to see how they look. Also in this category is Iowa (-14.5) who hosts Northern Illinois. I also expect big things from Iowa this year, as I picked them as the surprise winner of the Big Ten West. That said, losing in Week 1 to a solid MAC team is just the sort of thing Iowa likes to do. Finally, the sole Big Ten team to enter the week as a non-conference underdog is Maryland (+8.5), who gets the privilege of hosting Texas. My spreadsheet expects a much bigger blowout (+16).
In the long run, however, the two games this weekend with the most likely impact are the conference opener featuring Northwestern at Purdue (-3) and the not-quite-as-tradition laden-as-they-think match-up of Michigan (-1) at Notre Dame. (Fun fact: of the three other teams listed in this paragraph, Michigan is the team who has played the Irish the fewest times.) Interestingly, these games are also 2 of my 5 total upset picks for the week. Quite honestly, neither of these Vegas Lines make sense to me. As for Purdue and NW, the preseason magazines pretty universally had the Wildcats ranked about 20 slots higher than the Boilers. While the Boilers are at home and while I pegged Purdue as another potential surprise team this year, I would have to conclude that Vegas believes NW's starting QB is injured for this line to make sense. Either way, the losing team will be very much behind the eight-ball for the Big Ten West race. My calculations only give Purdue a less than 1% chance to win the Big Ten West right now and Northwestern's odds would drop from 10% to 5% is they can't escape West Lafayette with a "W."
As for the Irish and Wolverines, the preseason magazines had the two teams in a virtual tie as borderline Top 10 quality. If the two teams really are equally matched, you would expect the spread to be about 3 points for the home team (Notre Dame), which is exactly what my spreadsheet predicts. In contrast the spread was around 3-points for Michigan most of the summer, and opened this week as the Wolverines at -1. Since I always use the opening spread in my calculations and analysis (as it is a stationary and not a moving target), I am using that value. But, in reality, the spread seems to have moved almost immediately in the Irish's favor. Almost all tabulated computer prediction algorithms that I have found concur. Either way, the loser is certainly in a major hole if they have plans to make the playoffs (and of course they do). On the bright side, one of the two teams has to lose!
SEC
Similar to the Big Ten, essentially half of the SEC decided to partake of some sweet, sweet cupcake action this weekend, except they prefer their carb intake to take the form of FCS-level teams. Arkansas, Florida, Georgia, Mississippi State, Missouri, and Texas A&M all fall into the category, and South Carolina (-29.5) scheduled Coastal Carolina, a Sun Belt team who is so new to the FBS, the logos on their field are likely still damp. (It's the humidity). Also, Central Michigan is not likely to give Kentucky (-17) much trouble. But, the other six games are a bit more intriguing.
SEC doormat Vanderbilt (-6) is hosting Middle Tennessee State, and my math likes the Blue Raiders to cover, if not win. Trust me, Vandy, getting upset by Middle Tennessee is NOT FUN. But, it is the other 5 games that are the real must-see-TV this weekend as they all feature an SEC team playing another Power 5 team on a neutral (more-or-less) site. In Houston, Texas Tech will square off against Ole Miss (-1). Neither team is picked to be competitive in their respective leagues, but this game looks to be competitive, and my algorithm likes the Red Raiders in an upset. Charlotte will play host to a match between West Virginia (-8) and Tennessee, which my data suggests will not be as close as Vegas thinks. In Arlington, Miami (-3) will tangle with LSU in JerryWorld. If LSU is able to upset the Canes, it would be a major blow to a team that has a surprisingly reasonable shot at the playoffs. The defending champs, Alabama (-25.5) will travel to Orlando to face Louisville. Although Vegas thinks this will be a blowout, my spreadsheet predicts the Cardinal will cover easily. Not win, mind you, but just keep it under 20.
But the main event has to be a match-up of Washington vs. Auburn (-2.5) in Atlanta. With both teams entering the season in the Top 10 and both contenders in their leagues, this will be the first real heavy weight battle of the year. My spreadsheet also likes the Tigers, but only by the slimmest of margins (0.6 points).
ACC
Much like the Big Ten and SEC, the ACC has a handful of laughers on the schedule, sprinkled in with some interesting contests. Clemson, Georgia Tech, NC State, Pitt, and Virginia all open with FCS opponents, while Boston College (-18) and Wake Forest (-8) are double digit favorites against teams in the Bottom 40 of the FBS. Duke is -11 as a host to Army, and I mentioned above Louisville (+25.5) and Miami's (-3) big tests against Bama and LSU.
The three remaining contests feature two games that should be entertaining, if not impactful, and one that is clearly the biggest conference game of the weekend. In the non-impactful category, we have North Carolina going out to Berkeley to face the Cal Bears (-5) and Syracuse (-6) traveling to Kalamazoo to face Western Michigan. My spreadsheet does not pick an upset in either contest, but it likes WMU to cover (+2.4), and I am going to issue an upset watch for the Orange, as conditions seems favorable for the formation of an upset.
But, one of the biggest games of the weekend is the ACC opener featuring Virginia Tech at Florida State (-5.5). The Seminoles started the season last year as a consensus Top 5 team, yet injuries piled up early and they limped to a 6-6 regular season record. Virginia Tech is coming off a 9-3 regular season and some of my calculations suggest that the Hokies were better last year than we thought and might just be the surprise team of the ACC this year. My calculations also suggest that if the Hokie score the upset, they will have slightly better odds than Miami (49.5% to 46.4%) to win the ACC Coastal Division. However, a Florida State win could restore their confidence. Considering Clemson needs to travel to Tallahassee this year, one more win might be all it would take to shift the balance of the Atlantic Division as well.
Big 12
The Big 12 took a look at some of the scheduling in the other conferences this week and thought, "hold my beer." A full 60% of the 10-team Big 12 are facing FCS opponents, and 3 of the remaining 4 teams somehow managed to schedule Power 5 teams outside of the Top 50 in the previously mentioned Texas (-8.5) at Maryland, Texas Tech (+1) vs, Ole Miss and West Virginia (-8) vs. Tennessee contests. So, ironically, the highest ranked opponent a Big 12 team will face is C-USA favorite Florida Atlantic (preseason rank #36) who is travel to Norman to face the Sooners (-21). My math says that spread is a bit high (I peg it at -15.7), and my gut says that if you want to pick a crazy upset for Week 1... look here.
Pac 12
Due to the West Coasters generally dismal Bowl performance last year, I have demoted the Pac 12 to the last slot in my weekly analysis (Congrats Big 12!). As their first act of penance, they only put one FCS team on the schedule this week. That honor goes to Utah, who actually did win their Bowl Game last year, so that all seems fair. Honestly, there is not a whole lot to get excited about, as the Arizona schools, Oregon, Stanford, UCLA, and USC are all better than 2 TD favorites (although that San Diego State at Stanford (-15.5) game is a bit intriguing...), and Oregon State is close to a 40-point dog to the Buckeyes. The aforementioned UNC at Cal (-5) game should be watchable, and I suppose the Colorado (-6) vs, Colorado State game is always fun. I even have a predicted upset to watch for when Wazzou travels to Wyoming (+3.5).
But, when it comes down to it, the biggest game of the week is the previously mentioned match-up between Washington and Auburn. This game seems much bigger for Washington, honestly, as a loss would not just hurt their playoff chances, but it could effectively sink the whole Pac 12. Outside of Stanford or USC beating Notre Dame or USC beating Texas later in the season, there just aren't that many chances for the Pac 12 to gain a signature win. If Notre Dame were to also lose to Michigan, the entire conferences hopes of a quality victory would rest on the idea that Texas doesn't also suck. Good luck with that. Considering the conference went only 1-8 during Bowl Season, this is a game the Huskies really need to win, both for themselves, and for their conference.
Independents and Group of Five
We already talked about the Fighting Irish, so what about the Group of Five teams that could make a run at the New Years Six Bowls? In most years, Week 1 offers an interesting test for at least one of those schools. Well... not this year. I did mention the Florida Atlantic (+21) at Oklahoma game previously and other than that, unless you are really interested in Boise State (-10.5) vs. Troy, I suggest we just wait until next week.
That is all I have for now, so as always, Go State, Beat the Aggies!
Comments
Post a Comment