October 2nd, 2018 at 12:21 PM ^
This is what is called "Confirmation Bias". In that our biases against the shitty refereeing are now confirmed.
October 2nd, 2018 at 1:27 PM ^
But wait, there's more!
Using the data from https://www.ncaa.com/stats/football/fbs/current/team/694 , Michigan is #1 (or last, depending on how you look at it) in the FBS in opponent first downs from penalties with 20.
The FBS average is 8.92, with a 3.59 standard deviation, meaning Michigan is 3.08 standard deviations above the mean (a statistical outlier with a .001 probability assuming a normal distribution)
Despite this, Michigan is still in the upper half of the FBS in opponent first downs. This means other teams out there with worse defenses are potentially out on the field longer, leading to more chances for penalty first downs to happen. This led me to also look at Percentage of Opponent First Downs from Penalties. No surprise, but Michigan is #1 again with a ridiculous 23% of our opponent's first downs coming from penalties.
The FBS average is 9.67%, with a standard deviation of 3.76%, meaning Michigan is 3.55 standard deviations above the mean (0.0002 probability in a normal distribution).
fyi I'm not a stats wizard so be nice to me if I'm making bad assumptions. Thought this would maybe be worth making a thread for, but I have 0 points because I'm new and voting doesn't work anymore.
October 2nd, 2018 at 1:56 PM ^
3+ standard deviations from the mean!
This sort of anomaly will stretch the universe too thin.
This Big Chill infects us all
October 2nd, 2018 at 2:57 PM ^
Holy smokes. The picture you paint is ugly, but very nice work. Thanks for this awesome summary of our fan woe.
October 2nd, 2018 at 3:11 PM ^
Cincinnati is over 24%. Wash St. at 20%. So I wouldn't read too much into the normal distribution probability.
October 2nd, 2018 at 5:30 PM ^
Seems relatively normal to me
October 2nd, 2018 at 7:22 PM ^
So there's Michigan, a with a probability of 0.0002, and then another even less likely performance? What are the odds of both Michigan and Cincinnati hitting these numbers in the same time period?
October 2nd, 2018 at 7:44 PM ^
Yeah it's obviously not perfectly normal, I said that smart ass. I didn't say it was literally a .0002 probability. It's clear though that those two teams are far, far away from the other 127 teams for whatever reason, which is all I've been trying to show.
What're the odds that Michigan draws the fewest holding calls in the big ten (and it's not even close), while also having their opponents get the most first downs given to them by refs (and outside of one other team in FBS, isn't even close)?
It's low.
October 3rd, 2018 at 7:48 AM ^
Like I said, what do I know. Maybe I picked something up in grad school or 30 years of working in statistics?
But, yeah... I’ll take your eyes’ word for it.
Would one use an OLS regression model to predict this? Likely not, as it would allow for negative predicted values. Poisson regression, however, would be appropriate for predicting a discrete/ count distribution.
But, yeah, your eyes are probably right.
October 3rd, 2018 at 11:42 AM ^
+1 for Poisson.
Still, we can agree with the premise "Michigan gets boned by refs"
October 2nd, 2018 at 3:30 PM ^
This is a problem when your game is "drink every time an opponent gets a first down via penalty".
October 2nd, 2018 at 10:09 PM ^
Pretty sure that’s Poisson distributed, not normal.
But, what do I know?
October 2nd, 2018 at 4:11 PM ^
Yeah, the axes might as well be labeled: "Fuck Michigan/Fuck Harbaugh"
October 2nd, 2018 at 5:26 PM ^
....or axis, just sayin'.
October 2nd, 2018 at 10:48 PM ^
No, axes (plural). One labeled "Fuck Michigan;" the other, "Fuck Harbaugh."
Just sayin'.
October 2nd, 2018 at 9:38 PM ^
Guys, just to note - I updated my initial analysis for 2015-2018 week 5 (and added Iowa back in :))
October 2nd, 2018 at 12:23 PM ^
Now if a statistician wants to take this a step further they can evaluate whether the bias is more pronounced by in specific officiating crews. I.e. If the O'Neill crew is a lot more biased than the others Michigan should refuse to play in a game officiated by them.
October 2nd, 2018 at 12:26 PM ^
I'm guessing there's not a high enough sample size to determine bias among particular individual crews.
October 2nd, 2018 at 12:46 PM ^
Might still make sense to rule out an officiating crew even if the confidence interval is a bit high.
They could just officiate some other Big Ten game. No harm to that crew.
October 2nd, 2018 at 1:08 PM ^
It’s my analysis; and I did look at referees. Sample size is too small to glean anything useful. O’Neill was in the middle of all referees for Michigan games.
October 2nd, 2018 at 1:50 PM ^
It would be very useful also to separate the date into 2 parts: Big Ten referee called penalties, and all the rest. I'm guessing that Michigan would not be nearly as big an outlier in the latter data set. But it would need to be verified.
October 2nd, 2018 at 12:23 PM ^
Dang, even Pluto is more of a planet than us.
October 2nd, 2018 at 12:23 PM ^
Where is Iowa though!
October 2nd, 2018 at 12:25 PM ^
Iowa. Guy that posted it said they'd be somewhere in the middle and just forgot to include them
October 2nd, 2018 at 1:11 PM ^
Yup, I totally shit the bed on that one! Note to self, when you filter the teams in Tableau, maybe go off of a list and not by memory!
October 2nd, 2018 at 1:12 PM ^
To be fair, forgetting Iowa is a very easy mistake to make. In football, and life in general.
October 2nd, 2018 at 1:29 PM ^
relevant every four years or so
October 2nd, 2018 at 2:43 PM ^
Guess this year ain't their year.
October 2nd, 2018 at 12:35 PM ^
They are out harvesting corn, no time for stats.
October 2nd, 2018 at 12:43 PM ^
This is outstanding, and you won't get enough upvotes!
October 2nd, 2018 at 12:44 PM ^
South of Minnesota.
October 2nd, 2018 at 2:14 PM ^
HAHA....Nice one!
October 2nd, 2018 at 12:24 PM ^
That Michigan is not getting the holding calls that they should is expected, but Jeebus, seeing this chart is shocking. This is complete injustice.
October 2nd, 2018 at 1:10 PM ^
fake news
s/
although it would be interesting to see if this were disclosed to other places like a B1G office.
October 2nd, 2018 at 12:24 PM ^
Always worth repeating. People need to be aware of this.
October 2nd, 2018 at 2:08 PM ^
What makes it even more eye-opening is that for the last few years at least Michigan has featured one of the best, if not the best, pass rushes in the conference. By its very nature it should be above average in holds caused at the very least.
October 2nd, 2018 at 3:04 PM ^
Boggles the mind. All the other teams show a definite and significant correlation between holding calls against and sack rate. All bunched relatively close to the trend line.
And then there's M with the highest sack rate and an asburdly, nonsensically low holding against rate.
October 2nd, 2018 at 3:47 PM ^
we're so fast that they can't hold onto us
October 2nd, 2018 at 12:26 PM ^
Stunning. Just stunning.
October 2nd, 2018 at 12:26 PM ^
That's neat, but still nowhere near the coolest statistic/correlation ever discovered...
October 2nd, 2018 at 1:12 PM ^
this doesnt include fast times at ridgmont high in 1982
October 2nd, 2018 at 1:41 PM ^
Technically, he wasn't Nicholas Cage in that film. He's credited as Nicholas Coppola.
October 2nd, 2018 at 1:37 PM ^
Wow. Nicholas Cage really decided to tell humanity to fuck off in 2007....
October 2nd, 2018 at 2:28 PM ^
Yup. That was the year "Ghost Rider" came out, which was essentially his way of saying he hates every human on the planet...
October 2nd, 2018 at 12:27 PM ^
Does anyone have a statistic for holding calls that are declined?
If the defense gets to the quarterback and sacks him we would decline the holding.
Just want to see if that stat does anything to equal out the whole story.
With all that said I feel we're getting screwed.
October 2nd, 2018 at 1:12 PM ^
On average, teams defend about 70 plays per game (for 2017 offensive plays per game for 2/3 of D1 was between 65 and 75).
Indiana gets just under one hold per game called against opponents.
Michigan gets just over one half of one hold per game called against opponents.
Sample size is a SERIOUS issue with any attempt to analyze this. You really need a lot bigger data pool, both in terms of number of teams and number of years under consideration.
October 2nd, 2018 at 2:00 PM ^
Oh, sorry I missed the far right column.