Fan Satisfaction Index: Regular Season BBall Survey

Fan Satisfaction Index: Regular Season BBall Survey

Submitted by OneFootIn on February 24th, 2018 at 2:06 PM

Michigan's Men's basketball team ends its regular season with a win at Maryland. Having spent the fall polling fans about their feelings about football, it's time for our first fan satisfaction survey for basketball.

How do you feel about the Maryland game?
How do you feel about Michigan's regular season?
How do you think Michigan will do in the NCAA tournament?

Take the survey here:

For more information:

Introduction the Fan Satisfaction Index

Fan Satisfaction Index: Ohio State Results

Fan Satisfaction Index: Ohio State Results

Submitted by OneFootIn on January 13th, 2018 at 11:52 AM

Note: Sorry this is so late – work and the holidays conspired against me this year.

Sigh. Another regular season ends with a disappointing loss that could have been a win. Buoyed by a great game plan, the Wolverines jumped out to lead, made me break my promise not to have any hope whatsoever, and then the football gods took that hope away and crushed my heart. Again. Yeah, Harbaugh has things pointed in the right direction and the future is bright. But I live in the present and in the present I feel like shit (Edit: this goes double after the Outback Bowl – see part 2).

And so, evidently, do most of you. As I will explain below in just a bit, game satisfaction “without trolls” checked in at 27.7. This was almost identical to the Wisconsin game (28.8). This surprised me some given it was another loss to our biggest rival, though the Wolverines certainly played a better game than most people expected. A less optimistic take, on the other hand, might be that the Michigan fan base has become a bit numb from losing so often to the Buckeyes and that low expectations led to less anger and upset than is sometimes the case.

Figure 1. OSU Game Satisfaction

Season satisfaction (without trolls) also held more or less steady from last week at 36.8. In scientific terms this means the season was…not good. As I discussed last week, even if your rational self knew with great certainty that an 8-4 record was the most likely result of this season, you still felt like shit on Saturday. It turns out that expecting 8-4 and *experiencing” 8-4 are two totally different things. Sure the season probably would have felt worse had we expected to go undefeated, but losing is losing and no one likes it.

Figure 2. Season Satisfaction after OSU

Thus the regular season ended with satisfaction on a decided down note after the "Peters Resurgence."


Figure 3 Season Trends

Themes, Thoughts, Trends

Here Come the Trolls

The trolls found our survey. It’s the Internet so I knew it was bound to happen, but still. This is why we can’t have nice things. Of the 227 responses I logged for the OSU survey, I estimate that somewhere between 15 and 33 of them were our enemies – you probably know them as “jive turkeys.”

How do I know they were trolls? Well, if you rated both your game and season satisfaction as 100, as 15 people did, then I’m pretty sure you’re a Buckeye (or possibly a Schadenfreude Sparty) taking the survey for the lulz. Another 5 people rated their game satisfaction as 100 but with a strange variety of other season satisfactions. And another 13 people rated their game satisfaction as somewhere between 80 and 99.

Now, I’m sorry, but an actual living and breathing Michigan fan does not give this game an 80. Did you? If you are a real Michigan fan and you did, please let me know in the comments. Otherwise I have to assume you were high or live in Ohio, or likely both.

That said, after a long conversation with my scientifically inclined son, I realized that in the name of science we couldn’t just delete data, even Buckeye data. So in the interest of transparency and truth and the like, here is your satisfaction sensitivity analysis, under various troll identification parameters.

As you can see, there are enough trolls to make a difference in the results.

Table 1. Who’s Trolling?

Troll ID Rule Game Sat Season Sat

# Clean Responses

# Trolls
Assume no trolls 37.7 42.2 227 0
Game & Season Sat = 100 33.3 38.1 212 15
Game Sat = 100 31.6 38 207 20
Game Sat = 80+ 27.7 36.8 194 33






Another way to find the trolls is to use a simple scatterplot. As you can see, there is an obvious central cluster and then there are some obvious outliers near the maximums on each axis. These are probably your trolls. It’s even more obvious something’s fishy when you compare this data to the data from Michigan’s wins (which were unlikely to result in opposing fans filling out our survey). In those cases there just aren’t any fans adopting the 0/0 position – so I’m pretty confident we can rule out anyone who answered 100 on both counts.

Figure 4. Scatter Trolls

What I am curious about, though, is what you think the most reasonable cut off point is. Is there any way a Michigan fan gave that a 100 for game satisfaction? Or an 80? Maybe on the notion that the lads did their best and gave the Buckeyes all they could handle, etc., etc.?

The Road Ahead

I was going to point out how there was one more shot at redemption, a chance for at least a moderately optimistic ending on the season.

But since I’m writing this after the Outback Bowl I won’t bother.

Stay tuned for part 2 for results from the Outback Bowl and to see how other B1G fanbases fared this season.


Fan Satisfaction Index: Outback Survey

Fan Satisfaction Index: Outback Survey

Submitted by OneFootIn on January 1st, 2018 at 3:41 PM

Fan Satisfaction Index: Ohio State Survey

Fan Satisfaction Index: Ohio State Survey

Submitted by OneFootIn on November 25th, 2017 at 3:44 PM

Fan Satisfaction Index: Wisconsin Results

Fan Satisfaction Index: Wisconsin Results

Submitted by OneFootIn on November 22nd, 2017 at 11:40 AM

That sucked. It started as a hard-hitting Big Ten rock fight with Michigan giving Wisconsin all it could handle. Then it dissolved into another miserable and hard to watch affair as Peters went down with a concussion and the team lost its mojo. On paper Michigan did pretty well; Peters had his most promising outing despite a couple of mistakes and the defense kept Wisconsin’s high-powered tailback in check for most of the game. Heck, with a couple of breaks (thanks replay guys!) the outcome could have been very different.

But things went the way they did and from a fan’s perspective it mostly just sucked. This week’s game satisfaction clocked in at a whopping 28.8, down almost 50 points from the past three weeks, slotting in just ahead of the Penn State debacle.

Figure 1. Wisconsin Game Satisfaction

Unsurprisingly, season satisfaction also took a nosedive, dipping to 36.3 after camping out near 70 the past two weeks. At this point most fans seem to be grappling with the cold hard truth that Michigan is likely to end the season 8-4 and without any quality wins (in fact, without a win over a team with a winning record).

Figure 2. Season Satisfaction after Maryland

This brings up an interesting point about fan psychology. Before the season started a lot of analysts, including our own Ecky Pting, predicted Michigan would go 8-4 this year and would have trouble doing better than 9-3. Michigan was rebuilding on both sides of the ball, had tough games on the road against Penn State and Wisconsin, and beating Ohio State is always a challenge. In theory, then, fans should be relatively sanguine about going 8-4. Most fans, of course, are decidedly not sanguine about it. Many of them are losing their shit. The threads and comments this past week have been a mess.

There are many reasons for this psychosis. The most basic reason is that fans are not rational. Emotions don’t obey the laws of analysis and logic. Just look at Michigan Twitter during a loss if you doubt that statement. Feeling better than warranted after crappy wins and worse than warranted after tough losses on the road to the #5 team in the nation is just what it means to be a fan.

Somewhat more specifically, though, I think fans have problems setting expectations. They look at the fancy stats analysis that provides a rational and compelling case for an 8-4 prediction and then they immediately imagine all the ways in which Michigan could beat the prediction. Speight will be better than last year; the receivers are young but more talented; MSU will suck because they lost all those guys; we play OSU at home this year, etc. Pretty soon the fan is screwed because 10-2, not 8-4, has now become his or her emotional baseline for success. I know this because I am one of these people. I can know in my head that 8-4 is a sign of progress, but my heart will still bleed at the failure to go 10-2 (or better, really).

Themes, Thoughts, Trends

I Am Too Rational!

Okay, fine. Figure 3 provides some evidence that fans aren’t entirely irrational. The correlation between scoring margin and game satisfaction is quite high. We can explain 78% of variation in game satisfaction with just the margin of victory (or loss). In my regular season wrap up column I will use my somewhat more complete model to simulate game and season satisfaction scores for each of the other Big Ten teams – I have already done several of them and the variations are very interesting.

Figure 3 Scoring Margin and Game Satisfaction


The Road Ahead

Well, we’ve reached the end of the line. It’s the last chance for Harbaugh and the guys to pull our season satisfaction numbers out of the toilet. A win sends Michigan fandom into bowl season with confidence and boundless optimism about next year. A loss, especially a crushing loss, well, the less said about that the better.

Go Blue.

Figure 4 Season Trends



Fan Satisfaction Index: Maryland Results

Fan Satisfaction Index: Maryland Results

Submitted by OneFootIn on November 16th, 2017 at 10:54 AM

Since I live just outside D.C. I was lucky enough to catch last week’s game in person with my son. I’m pretty sure Michigan fans outnumbered Maryland fans, if not at kickoff then for sure by the second quarter after the drubbing was in full effect. It ain’t the Big House, but Maryland’s stadium is nice enough and the whole parking/getting in/getting out aspect was about as easy as I’ve ever experienced. Plus I got to sing the Victor's at a game, with actual Michigan fans, for the first time in about a decade. Overall gameday satisfaction level: high.

Harbaugh called it Michigan’s most complete performance of the season. I, on the other hand, gave it a “meh.” So, apparently, did the fan base, which gave the game a solid 73.5, which ranks fifth among Michigan’s wins, just behind Rutgers but well above Air Force, Cincinnati, and Indiana. Harbaugh is undoubtedly looking at the more important elements of the game here, but from a fan perspective the 73.5 seems fair. Michigan crushed a team they were supposed to crush, but didn’t look all that dominant for much of the game. The defense gave up more yards than expected to Maryland and its Nth-string quarterback, the offense looked good in spurts but fell short of exerting total dominance, and Quinn Nordin, #collegekicker, missed an easy one.

Figure 1. Maryland Game Satisfaction

Nor did the fans find much in the Maryland game to move the needle on their assessments of the season so far. Beating another cupcake this late in the season just isn’t going to change anyone’s opinion of the team and it isn’t going to erase the bad taste of losing to MSU and PSU. For that reason, even though it felt pretty good to thump Maryland, season satisfaction for the week checked in at 69.9, up less than a point from last week.

Figure 2. Season Satisfaction after Maryland


For the third week in a row game satisfaction outpaced season satisfaction. But one way or another Figure 3 is going to look very different next week…

Figure 3. Season Trends

Themes, Thoughts, Trends

Calibrating Our Instruments*

As good scientists it is critical that we be sure that our measurements are measuring what we think they’re measuring. In the context of our project here, we’d like to know that when Person A says a game gave them a satisfaction level of 80 that it means the same thing as when Person B gives us an 80.

Let’s take Netflix movie ratings as an example of this problem. If your family is like my family, some of you have discerning minds and impeccable taste, and rate movies accordingly, while others…do not. I, for example, give okay movies 3 stars, good or very good movies 4 stars, and I reserve 5 stars for things that are incredible. My daughter, on the other hand, hands out 5 star reviews like Halloween candy. Needless to say no one should pay any attention to our Netflix account’s overall ratings of movies. We would obviously like to avoid this problem with the fan satisfaction index.

Like Netflix, I have given very little in the way of guidance to respondents as to what they “should mean” when they answer a certain way. Other than telling them a zero is the worst and a 100 is the best, I figure your average MGoBlogger can handle scoring their satisfaction in a way that most folks will understand. We all have plenty of experience with tests that are scored from zero to 100, and most people have taken an infinite number of other surveys; there is good reason to expect that the hive mind will produce a reasonably tight consensus around what any given number means.

On the other hand, since I’ve never asked anyone to explain his/her response, I have had to assume that my interpretation of the hive mind is more or less the standard interpretation that anyone would give. But hey, I could be wrong. When I was surprised at how satisfied the fan base was about beating Rutgers, for example, I might have simply been misinterpreting what a 73.9 meant to the typical respondent. Maybe a 73.9 is really code for “adequate but not particularly invigorating victory over a crappy opponent.” If that’s the case, not only did I whiff a bit interpreting fan reactions to Rutgers, but my interpretation of other games should also be amended.

Here’s another interesting thing about trying to measure attitudes with surveys like this: most of the numbers on the scale are seriously underutilized, while a few more way overrepresented. As you can see in Figure 4, some numbers serve as something like a cognitive anchor. People are a lot more likely to answer 60, 65, or 70, than 51, 63, or 77.

Figure 4 Bumpy Data: 1875 Game Satisfaction Responses through Maryland


In fact, people are so much more likely to use certain numbers that it makes you wonder sometimes why anyone would use a 0 to 100 scale in the first place. Table 1 shows what I mean. The even numbered “primary anchors” (10, 20, 30, etc.) made 47% of all responses, the odd numbered “secondary anchors” (15, 25, 35, etc.) made up 33.5% of all responses, but the numbers adjacent to the anchors have been woefully under used. Even sadder: thirteen lonely numbers haven’t been used even once so far!

Table 1 Anchoring Effects and Survey Responses

Anchor % of all responses Anchor Adjacents % of all responses
0 2.13 1 0.43
10 2.61 9 & 11 0.05
20 2.61 19 & 21 0.1
30 2.4 29 & 31 0.0
40 3.31 39 & 41 0.05
50 3.73 49 & 51 0.7
60 4.91 59 & 61 0.16
70 8.37 69 & 71 0.85
80 11.52 79 & 81 0.7
90 4.59 89 & 91 0.95
100 .85 99 0.11
Total 47.03   4.1

Fun stuff.

In the spirit of getting it right, then, this week is Calibration Week. In the comments, tell me what you think the numbers mean. What’s a 50 to you? What’s a 75? How did you figure out what number to use?

* Get your minds out of the gutter. This is a family friendly site.

The Road Ahead

To Wisconsin we go…time for another test of our fan satisfaction system. This could get interesting. The game is at Camp Randall, a horrendously difficult road test for any team under any conditions, and the Badgers are favored by 7.5. Wisconsin’s offense has issues, but its defense has been stifling – not a great omen given Michigan’s aggressively mediocre offense this season. And yet…I’m sensing an awful lot of hope, even maybe the stirrings of expectations that Harbaugh and Co. have been keeping their powder dry and will finally unleash the Messiah’s true firepower in Madison to lay waste to the already frozen tundra. We shall see.






Fan Satisfaction Index: Minnesota Results

Fan Satisfaction Index: Minnesota Results

Submitted by OneFootIn on November 9th, 2017 at 9:22 AM

Is beating a mediocre Minnesota team with zero offense really worth an average game satisfaction of 78.2? I mean, I love the Little Brown Jug as much as the next guy, but I think I was 10 the last time I was actually excited about seeing us win it. But the fans loved it, rating their satisfaction with the Minnesota victory higher than any other game this season except Florida.

Figure 1. Minnesota Game Satisfaction 

Last week we tried to explain the exuberant reaction to the Rutgers game as a response to the arrival of the Messiah. That line won’t fly this week, as Peters turned in a solid but decidedly un-Messiah-like performance, completing just 8 of 13 throws for 56 yards.

If you want to stick with a player-based explanation, you can point to the double-barreled running attack of Higdon and Evans, who torched Minnesota for 377 yards and multiple explosive touchdown runs. And no question, for many of us the clear improvement in the ground game has been invigorating; watching Higdon and Evans blow past Gopher defenders was fantastic.

But perhaps a more compelling explanation for fan satisfaction this week is the fact that Michigan exceeded expectations. Not only did the running game go off unexpectedly, but Michigan also beat the spread by 7 points, its greatest margin against the spread this season other than – you guessed it – Florida. As Figure 2 shows, the R-squared between performance versus expectations and game satisfaction is .79.

Figure 2. The Expectations Effect 


Buoyed by consecutive (and satisfying) wins, season satisfaction crept up another notch this week to an average of 69.1, its highest mark since the Purdue game, when satisfaction was 76.7. Like last week, game satisfaction outpaced season satisfaction – fans are still not completely over the two losses, but I’m a little surprised that the season satisfaction rebound has been so significant. It’s also interesting how tightly bunched fans’ feelings are this week – the standard deviation in season satisfaction this week was just 11.6.

Figure 3. Season Satisfaction after Minnesota

Themes, Thoughts, Trends

Another week, another bullet in the data chamber, and the seasonal trend lines get another tick longer. So far game satisfaction has averaged 69.2 across the 7 wins, and a meagerly 22.4 for the two losses.

Figure 4: The Season So Far

The Road Ahead

As Maryland continues to shed quarterbacks, Vegas has pegged Michigan a 15 or 16-point favorite over the poor Terps. It should be another chance for the running backs to go ham and hopefully Brandon Peters gets a few more live-fire opportunities to prove he is the Anointed One. An upset on the road would undoubtedly send satisfaction plummeting. Another hamblasting should see season satisfaction sneak even further up before the next real test in Madison.








Fan Satisfaction Index: Rutgers Results

Fan Satisfaction Index: Rutgers Results

Submitted by OneFootIn on November 2nd, 2017 at 8:41 AM

Is irrational exuberance blossoming in Ann Arbor?

In a 1996 speech about the dot com boom Federal Reserve Chairman Alan Greenspan popularized the term “irrational exuberance.” The market was overheating he argued, and he warned that all market bubbles must eventually burst. And as we know, after a dazzling few years, the dot com hype eventually gave way to the cold harsh reality that businesses actually have to, you know, make a profit.

Brandon Peters came in for a struggling John O’Korn on Saturday and immediately made Michigan’s offense look more competent. Peters himself looked calm and in command, completing passes to nine different receivers and recording his first touchdown pass (and just as importantly – not throwing any interceptions). His presence on the field clearly energized his team. When Peters entered the game halfway through the 2nd quarter it was tied at 7-7. By half the Wolverines were up 21-7. Peters’ second half performance was drama free, allowing Michigan’s running game to close out the victory. Clearly the Messiah has arrived.

Skeptics have pointed out that Peters was playing Rutgers, was not asked to do much, showed some signs of hesitation out there, and that even if he is the next Messiah, there will be significant growing pains along the way. Others have noted that O’Korn looked like the answer after Purdue, only to be revealed as a false prophet. All bubbles, they seem to be saying, must eventually burst.

Maybe so, but after Saturday’s game fans were in no mood for such logic. Brandon Peters’ performance on Saturday not only satisfied the demands of growing chorus of fans calling for a quarterback change, it also seems to have reignited hopes for a season many had written off as lost. The result: both the game satisfaction index (73.9) and the season satisfaction index (62.9) are at their highest levels since Purdue despite the fact that Michigan did not in fact cover the spread. Interestingly, there was no difference this week between hot takes and cold takes. It just felt good.

Figure 1. Rutgers Game Satisfaction


Figure 2. Season Satisfaction after Rutgers

Michigan has clearly entered the roller coaster portion of its season. Figure 3 shows the whiplash fans have been through the last month.

Figure 3: The Season So Far

Themes, Thoughts, Trends

Rational Exuberance?

Looking back at Figure 3 you will note that this is the first week where the game satisfaction level was higher than fans’ satisfaction with the season as a whole. I think this reflects at least two things. First, many fans started the season with very high expectations, which was reinforced by the win against Florida. That, combined with a series of underwhelming wins, saw season satisfaction continue to outpace game satisfaction even as it dropped after losses to MSU and PSU.

Second, as jazzed as fans may be about Peters, even the Messiah can’t undo the losses already suffered. Nor are most fans ready to believe – yet – that Peters can deliver them to the promised land (victory against the Buckeyes) this season. This week’s results might just reflect a sort of rational irrational exuberance. Irrational hopes about the future with Peters mixed with lingering ennui and doubts about beating the Buckeyes. Sounds like a fan to me.

The Road Ahead

Vegas has Michigan a 14.5 favorite over the Gophers this week. I’m holding my breath to see what the Messiah looks like in his first start. You can feel the fan base holding its breath. If Peters fires two or three touchdown passes will fans lose their sh*t and start promising victories over Wisconsin and Ohio State? How do folks feel if Peters looks human and Michigan pulls out a close one? And God forbid, do the walls cave in if Peters looks like a freshman and Michigan stumbles again under the lights? We shall see.




Fan Satisfaction Index: Rutgers Survey

Fan Satisfaction Index: Rutgers Survey

Submitted by OneFootIn on October 28th, 2017 at 3:02 PM