Well, that happened. Pretty much all the ingredients for a massive dip in fan satisfaction were present on Saturday:
- Heartbreaking loss to a rival who had no business being any good this year? Check.
- Michigan favored by double-digits? Check.
- At home, under the lights, on national television? Check.
- Horrendous, nigh-unwatchable performance by the offense? Check.
- Highly questionable coaching decisions and play calling? Check.
- Mind-numbing displays of God’s anger at Michigan in the form of turnovers and bad luck? Check.
- Sinking feeling that this portends terrible things for the future? Check.
The loss came on top of rumblings and concerns about whether this team is really ready for prime time. Had the team taken a step forward during the bye week and paddled Sparty, the narrative would have shifted decisively for the positive. Instead, the disappointing loss did just the opposite, igniting complaints about Harbaugh and likely signaling an end to the honeymoon phase of Harbaugh’s tenure in Ann Arbor. It’s amazing how quickly fans have moved the goalposts. When Harbaugh came to town no one imagined Michigan would win 10 games in his first year. This week, the same folks ready to canonize Harbaugh are questioning his coaching ability.
Figures 1 and 2 provide our first look at how the fan base feels after a loss. It ain’t pretty. According to our respondents, the loss not only felt bad in and of itself, but it also pretty much destroyed the satisfaction fans were feeling about the season up to this point. The average game satisfaction plummeted to 20.8 – down from 76.2 after the Purdue game. The season to date score averaged just a 40.5 – also down from about 76 last week.
Figure 1: MSU Game Feelz
Figure 2: Post-State Season Feelz
The loss also appears to have undermined confidence in the future. Asked to assess Michigan’s win probability next week against Indiana, Michigan fans averaged just 66.3%, considerably lower than the 78.4% S&P had projected during the bye week (but more in line with the new Vegas spread, which is now just 5 points).
Sadly, given how bad Michigan’s offense looked, one has to assume that the 66.3% is only that high given the confidence fans still have in the defense, which acquitted itself extremely well once again. It will be interesting to see how much fan satisfaction manages to rebound when Michigan wins next week. I don’t think any of us want to see what happens if Michigan finds a way to lose to Indiana. I’m not sure numbers even go there.
Themes, Thoughts, and Trends
Recency Bias and Fan Satisfaction
One look at fan evaluations of the season after the State loss provides a powerful suggestion about the importance of the last game played in the fan’s mind. Fans, more than most people, seem susceptible to the “what have you done for me lately” disease. Of course, it’s also possible that rivalry games are just much more important than other games and that losing to State ruined any good vibes folks were having to this point.
So I went back to the bye week data. I ran a regression model using people’s ratings of season satisfaction so far as the dependent variable and their ratings of the first four games as the independent variables. In English, that means the result of the regression tells us how much influence each individual game had on people’s evaluation of the season.
The results were fascinating. The adjusted R2 was .49, meaning that we can explain roughly half of the variation in people’s assessments of the season by knowing their assessments of the individual games. Not bad. But the kicker was that only one variable turned out to be statistically significant: the Purdue game.
I’m not sure that says much for the mental acuity or mental health of the average fan, but it is evidence for the recency effect. My tentative conclusion at this point is that rivalry games have an outsized impact on fan assessments, but so does the last game played. And right after a rivalry game, those two factors combine to produce a massive effect on fans (And it probably explains why Brian and Ace complain so much about their Twitter feeds after games like this one).
The Season So Far
We riding the roller coaster now, folks. As usual, the columns represent game satisfaction; the line shows season satisfaction ratings.
Figure 3: The Season So Far, Summarized