Bad Betting Advice, Week 6

When it comes down to the most basic description of me as an MSU fan, the word that I always come a back to is “optimistic.” It is not that I believe that MSU is going to win every single game that they play. It is more that I usually try to think about how MSU might be able to win a game and how likely that scenario is to occur. I find that a study of historical statistics is a great way to take an unbiased look at the situation. For MSU’s contest in Columbus this weekend, I find this exercise to be particularly useful.

So, let’s address the bad news first. Ohio State looks like a juggernaut. It is possible that they are the best team in the country right now, and MSU needs to face them in their house. That’s not great. As for the numbers, the picture that they paint is not exactly encouraging.

To put things into perspective, in 2015, MSU was +13.5 at OSU. In 2018, MSU was +13.5 at Penn State. In 2017, MSU was +10 vs. Penn State at home. MSU won all 3 of those games, but those are the biggest home and away upsets for MSU back to 2001. MSU opened +17.5 in some places, but it rose so quickly that the source I use says it opened at +20. My spreadsheet is even less optimistic than that. It has MSU at +27.4 while the FPI is slightly more optimistic and has MSU at +18.4. Either way, if MSU pulls this upset, it will be the biggest MSU upset since MSU beat #1 OSU in 1998 in Columbus.

But, things are perhaps not quite as bleak as they seem. Teams that open around +20 actually do win straight-up around 8% of the time and this magnitude of upset does occur roughly 4-5 times in any given football season. We have observed three upsets of a similar or great magnitude already this year:

• GA State over Tennessee (+25.5)
• Kansas over Boston College (+18.5)
• San Jose State over Arkansas (+20.5)

So, while an MSU upset would be a huge deal, it is not crazy to imagine that it could happen. At the end of the day, college football has far more variability than most casual fans realize. Some of my past analysis has shown that the standard deviation of the final margin of victory relative to the spread is almost exactly 2 TDs. You can see that every week when I plot the final margins of victory versus the spread. In other words, an MSU victory would be a result that is only 1.5 standard deviations from the mean. That is unusual, but not exactly rare. In other words, there is certainly a path to an upset, even if it is a narrow one.

So, what might that narrow path look like? Basically, MSU needs to play very, very well. MSU needs to bring its A-game or A- game. At the same time, OSU probably needs to be a bit lack-luster. They need to bring their C or C- game. In past years, OSU has thrown out a stinker performance a handful of times a year. If OSU happens to have one of those types of games this weekend, then MSU will have a chance to win.

Specifically, I would say MSU needs to contain OSU’s run game and force Fields into mistakes. MSU probably needs to win the turnover battle by at least 2. MSU also needs to avoid the mistakes that have plagued the team since the Tulsa game. Specifically, MSU needs to minimize penalties, drops, bad throws, missed tackles, fumbles, coverage busts, and special teams errors. Simple, right?

At the end of the day, MSU needs to view this game simply as an opportunity. Despite the early loss to Arizona State, it is true that all of MSU’s goals are still in front of them. If MSU still has the desire to go to Indianapolis in early December, we are almost certainly going to need to win this game. If that does come to pass, then the discussion just might turn to the idea that if the officiating had been better in the ASU game, then MSU might be a playoff contender... If only.

As is my tradition, in order to put this data in perspective, here is my weekly preview graph for Week 6, based on my algorithm’s projections relative to the opening spread:

For the first time this week, I have removed the influence of the preseason rankings. As a result the correlation to the spread is a little weaker. While some teams have played 5 FBS teams, some others have played only 3 FBS teams, 1 FCS team (the results of which I ignore) and had a bye week. So, the data is still just a little bit sparse. It will get better from here on out.

For comparison, here is the same plot, using ESPN’s FPI’s projections.

I already mentioned above the detailed break down of MSU’s line and the probabilistic ramifications. The only thing that I will add is that my computer spits out a projected final score of 41-13, OSU. The over/under for the game is right around 50 points, so my model does like the over and, sadly, it likes OSU to cover. That said, both the FPI and S&P+ have MSU barely covering the current +20 line. I think that counts as one of the nicest things ESPN has said about MSU at least 2 years.

As for my recommended bets, the newly found freedom of my algorithm from the bounds of the preseason rankings has caused it to make all sorts of bold picks this week. A grand total of 14 games fall into the regime of a recommended bet, with one more bet from the FPI for good measure. The recommended bet table is shown here, in order of the strongest signal to the weakest:

If you are looking for a 2-team parlay this week, my math likes Penn State and Louisville to cover. I would be lying if I said I have a lot of confidence in these picks, but I think it is best to stick to the program and see how this plays out...

As for upsets, for similar reasons my model has a lot of picks. Those are summarized below:

Notable here are the UCONN over USF and Oregon State over UCLA. Those seem like bold picks. Finally, it seems that we have come to the point of the season where the FPI is mostly parroting out the Vegas line and in fact only has one upset pick. While that may have some value in projecting future lines, it really doesn’t add any value once the lines for each week are announced. Finally, I should note that my simulation of the week suggests that we will see 12.2 ± 2.9 upsets. Here’s to hoping one of them occurs in Columbus.

Also note a certain team for Ann Arbor is on the upset list as well. While Vegas and even some of the other computer models (such as the FPI and S&P+) still seem to think that Michigan is the better team, my spreadsheet is significantly more skeptical. Dropping Michigan’s preseason Top 5 ranking from their resume resulted in the Wolverines falling down to #57 in my power rankings while Iowa is up at #15. As such, my calculations call for a 11.5 point road win for the Hawkeyes as well as a strong recommendation to take the Hawkeyes to cover.

While I suppose it is possible that UofM is suddenly going to play like a Top 20 team, they literally haven’t done a single thing this year to warrant the assumption that they are even a Top 50 team. Sorry, but beating Rutgers silly doesn’t qualify. Recall that even in 2016, MSU beat Rutgers 49-0. It seems that Vegas and even some of the computer models are being sucked in by the fight song, the helmets, and the recruiting rankings. As for me and my computer, they are going to need to show that they can beat a team with a pulse, and I frankly doubt that we are going to see it this weekend. If Michigan fans were upset by the drubbing they received on the road in Wisconsin, losing by double digits to Iowa at home would be even worse.

That is all for now. Check in after the games to see once again how bad (or good?) my advice was. Until next time, enjoy, and Go State, Beat the Buckeyes!