You ARE freaking out, MAN. My battle with the 3 or 4 still equals 4 formula.

Submitted by jfs52 on

The 3 or 4 still = 4 formula or "You ARE freaking out, MAN."

 

I can't belive so many people are tired of speculation about the upcoming season.  It's the most exciting (and UNKNOWN) start of a season in quite some time.  Besides, what do you people do at work?  If the answer is "not speculate about sports" then what is it you do for a living that's so fascinating? 

 

Anyway, I will be honest, I have been freaking out (man), but in an attempt not to I thought back to a formula that my family has been using to calculate Michigan wins for a while, and actually under Lloyd Carr (and actually for Bo too, come to that), it was fairly accurate, except for the last couple of years, but those were very up and down and unpredictable, and even had names like the Year of Infinite Pain.  You can't predict a year with infinite pain.

 

The formula is the 3 or 4 still = 4 formula, and requires the steadfast refusal to listen to all the if's and a refusal to listen to a lot of analysis- or the discipline to discount it.  The theory is that in college football, tradition is still very important, and except when Charlie LLLLLLLLWLWWeis is ruining it, teams are really surprisingly consistent despite other factors.  In this theory tradition wins out, and what usually happens will happen again.

 

So now I am wondering out loud if things will really change this year, or if a different system will yield really similar results?  We will see, and to do it I will introduce the formula.  To use this, the analysis is REALLY simple.  Is the team:

 

quite good? Give it a  1/2 (still equals 2)

pretty good, a  2/3 (still equals 3)

good, but a little disappointing for Michigan: 3/4 (still equals 4)

 

That's all the analysis that's required, and these are the ONLY choices.  What the numbers say is this:  If we win the lower number, we will be overrated and lose our bowl game, giving us the higher number of losses anyway.  If we lose the higher number we will be mad, get a worse bowl opponent and beat them, keeping the number of losses the same.  Like I said, the last few years were weird, but take 2004, Chad Henne's freshman year.  This was probably a "pretty good" team.  During the season they only lost twice- maybe playing just a hair above themselves.  But they lost (albeit by a hair) to a good Texas team to get 3 losses.  If that team had lost 3 regular season games, though, I submit that we probably would have won that bowl game.  Hence 2 or 3 still equals three.  The reasoning behind only 3 choices is also simple- the vast majority of our seasons have had 2, 3, or 4 losses, and so that's where we start.

 

This analysis obviously sacrifices something up front: you can't predict years that are weird (and this admittedly could be one, by the way).  It is a system that admittedly aims to be only USUALLY right.  It wouldn't have predicted the national championship year, it wouldn't have seen last year's injuries.  But I would argue that those circumstances are usually unpredicatable anyway: if you HAD known about the circumstances, how would you have predicted least year? Knowing Not full-strength Chad Henne and remembering that we had lost 3 defensive superstars, you might have said this team was only "good", so 3/4 still equals 4, and you'd have been right.  Ok, so a LOT of circumstances come out in the wash, but remember we take for granted a lack of Henne.  You say 4 losses and we win the bowl game.  We shouldn't have beat Florida, true, but we also shouldn't have lost in The Horror, so like I said, it comes out in the wash sooner or later.  If that team had only lost 3, and we had had to play "up" a game then we almost certainly lose to Tennessee or USC.  It's one of those things- if you try to pick MOST of the games and seasons right you can, but somehow if you try to pick them all you get more wrong, just like the NCAA basketball tournament bracket where you thought about it for a long time and then all of a sudden you have too many 12 seeds winning games.  Your five minute bracket was a lot better.  This is the same principle.  Think about it this way: who is more right, the person who says the Lions will not go to the playoffs ever, or the person who tries to pick the year they get there?  It's the former, of course.     

 

This analysis is strong in its simplicity and its ability to either ignore factors that will be inconsequantial or combine with other factors to cancel out.  Football analysis is hard, and it's fun, and I love it, but really, I think we tend to overthink it.  In almost any given year Michigan will be good, Indiana will suck, and Michigan State will blow it at some point.  This analysis is true for like 95% of the college football seasons ever.  Like I said, the great thing about this system is it makes up for things coming out in the wash.  Let's take 2003.  I actually argued at the time that despite all the seniors this team was only "pretty good" because I didn't quite trust the defense or John Navarre.  And our competition that year was quite good- I think that people would have said that losing two out of three to Notre Dame, Michigan State, and OSU was totally possible.  Well, actually we won all of those games, but then crapped out at Oregon and Iowa.  Still though, the 2/3 formula was right as we lost to a better USC team.  I think if that's a 3 loss team during the season that they tear up the Capital One Bowl to stay with 3 losses.       

 

If you want to get a little more complex and you want to be able to predict an undefeated (or 5 loss) season then you can add a season modifier: +1 for a good chance something horrible happens, -1 something wonderful happens.  But it's hard to quantify that, so I don't usually.     

 

So even though my natural tendency is to freak out and say we have new (or even no) QBs, and a crappy OL, and that we will all be put to the sword, and the stadium will burn with red flame, and we will lose every game, the formula suggests only 4 losses.  (And to imagine a day when we said "only 4", but that's just me yearning for Bo again...)  Anyway, the 4 losses- this year we have to be a 3/4, not better, and I'd even say there is the potential +1 for something horrible (Notre Dame plays one good game this season).  My prediction is we lose 4 during the regular season (Utah, MSU OR ILL, PSU OR WISC, OSU), get forgotten about, and pulverize someone in a crappy bowl, like the Aloha Bowl or something.  But I think surely if we somehow lose only 3, which would be very (VERY) good result I think, in a bowl game we have to play some team from the SEC east that is almost certianly overrated, but also almost certainly better than us.  So four losses either way.       

 

Now, sometimes you have to listen to the analysis.  Sometimes everything is not business as usual- sometimes you can't just have your head in the clouds and wish something away.  I hope that's not this year, because when I sit down and do the analysis and think that with no QBs, no OL, and no LBs we suck suck suck.  So I sit down and try to trust the formula.  Still, we all know what happens with willful ignorance- "Rodents of unusual size?  I don't think they exist..."   

Comments

ShockFX

August 29th, 2008 at 4:18 PM ^

So, this system works, except when it doesn't?

"except for the last couple of years"

"except when Charlie LLLLLLLLWLWWeis is ruining it"

"you can't predict years that are weird "

Michigan, on the whole averages 3 losses. So if you always guess 3, you can say that 2 or 4 loss seasons were kind of right, and since we never have 1 loss seasons you can ignore those. That leaves 0 and 5 loss seasons. Both are completely rare and you already threw those out with "You can't predict a year of infinite pain" and "It wouldn't predict a National Championship".
Also, if a prediction accounted for future unknowns it wouldn't be a prediction, it would be prescience. And in that scenario you would hopefully be able to prevent the injury from happening by sharing the information like the Twilight Zone or whatever the hell that USA tv show is. Quite possible one of the worst things I've ever wasted my time reading.

KRK

August 29th, 2008 at 5:37 PM ^

I typed a really long response initially, but your post was so confusing that my reply started to confuse me. I don't think your formula works mainly because it isn't a formula. No formula will work for this year. If you want an indicator on the season go with this. We will win somewhere between 6-8 games this year with a margin of error at +/- 4 games. That should be the only "formula" anyone uses.

Note: You can't just throw out the stats that prove your formula wrong. Like "not being very good proves this wrong, so we won't use it." That probably won't fly with the Scientific Method.

Tacopants

August 29th, 2008 at 7:22 PM ^

I read this over a couple of times, I still don't understand half of it. I don't think you can call anything a formula if you're going to throw out the anamolies.

All I can basically glean from this is that Michigan has a bell curve distribution on wins in a 11/12/13 game season with the median being at 9 and a SD of 1. You'll have most of your season fall between the 8-10 win range, 95% will fall in the 7-11 win range, and so on.

"This analysis is strong in its simplicity" - incorrect, if you wrote that much about it, it is not simple. Here's simple: 75% of the time, I predict Michigan will win every time. (Or go on to 9-3 seasons)

Blue Durham

August 29th, 2008 at 9:54 PM ^

and not to pile on, and I only bothered glancing through your diary, but past trends don't really apply to this season. There are just too many variables, starting with the coaching change (duh, but this one was incredibly sweeping in its bredth and philosophy) to extrapolate from past season.

This season, to quote an academy award winning movie "fasten your seatbelts; it going to be a bumpy ride."

chitownblue (not verified)

August 30th, 2008 at 2:54 AM ^

I understand that I have a reputation of being a dick around here. But, personally, I don't see a problem in disagreeing with someone else. To whit: this post is a grammatical and logical holocaust.

I barely have an idea of what you're attempting to say, and what I can divine makes no logical sense, at all.

Also, I'm struggling with your assertion that we will lose to either Illinois OR MSU, Wisco OR PSU. The outcome of one game is not dependant on the outcome of another. Losing to Illinois does not guarantee a win vs. MSU, and a win against Illinois doesn't guarantee a loss. Same with Wisconsin and PSU. Why phrase it in such a way?

"Michigan football ends this way, aside from these times that it didn't - that I've chosen to ignore." No shit?

KRK

August 30th, 2008 at 7:20 AM ^

Who around here has thought you were a dick?  That's like saying I'm a sarcastic smart ass.  I do agree with you on this.  I can make any theory or formula if I throw out the times that it doesn't work.

wildbackdunesman

August 30th, 2008 at 7:43 AM ^

I think that you are forgetting that Michigan under the "Bo-Mo-Carr" era, was absurdly consistent. In fact we never had a losing season during that stretch. No program in college sports anywhere has been that consistent and unfortunately it can't stay that way forever.