The Monty Hall Problem

(Hello world. I am Alan. I used to be Jeff’s RA in the dorms when he was in college. Jeff’s been busy with real life so I asked him to give me access to the blog to help fill in.)

The MSU faction, ever since Sunday, has done everything they can on my Facebook wall to attest to the fact that they are more deserving to play in a BCS bowl than Michigan. Columnist Drew Sharp, who still believes that a Big 12 championship game still exists (it doesn’t), and others argue that common sense dictates that the winners of both B1G division should go on to the two most prestigious bowl games. Thus, MSU getting snubbed out of going to the Sugar Bowl speaks to the abomination of the BCS system as well as the fact that biased towards Michigan’s elitist stature in the college football world.

Unlike Jeff and Mike, I grew up in Michigan. So I must tolerate this talk but I do want to take this opportunity give a reasonable response to their grievances.

In elementary probability, there is classic example called the “The Monty Hall Problem”. Crudely explained, it demonstrates that intuition, or common sense, can often fail in mathematics.

The restless MSU faction, along with other columnists like Drew Sharp, says that anyone with any common sense can see that the BCS has screwed over teams like Oklahoma State and most importantly all those who bleed green and white. Sorry, but you’re barking up the wrong tree. It is the human polls, not the BCS, which screw you over, if anything.

I will start by addressing the root of the problem- on Sunday MSU was not eligible for a BCS bowl invitation. Thus, if we still believe BCS has screwed everyone, then the fault must lie in the computer rankings.

To quickly summarize, BCS rankings is a weighted average of both human polls (two-thirds of the overall weight) and 6 computer formulated rankings (each one-eighteen of the overall weight)

Without getting into too much detail*, the averages of the six BCS formulas acts similarly to college’s basketball’s RPI- it rates the entire body of work of a football team. Looking at MSU’s nonconference schedule (Youngstown St, FAU, CMU, loss to ND), it is easy to see why the computer rankings punished MSU. Honestly, if this is college basketball, MSU would be a bubble team with an iffy resume. That is not the problem.

MSU fell victim to the reality that it is counter-intuitive for anyone that understands competition to vote a team higher than the previous week when it loses. Other teams, who performed poorly last Saturday, along with MSU, also dropped. Thus, M moves up by a process of elimination in the human polls, which is then weighted as two-thirds of its BCS rankings. If there is any blame to be passed around, it is the structure of the AP and USA polls in relations to the BCS rankings. It is the human polls that cost MSU more than anything else, not the BCS. As the matter of fact, problems as such has exists decades before the invention of the BCS (most notably, 1973 in the Big Ten).

So blame Art Briles, Doc Holiday, and Kevin Sumlin for voting MSU #21 on their ballots (by the way Nick Saban voted MSU #17). It isn't about the fallacy of the BCS system, nor is it about Michigan.

Complaints are much more valid when it is directed at the actual problem, or if Drew Sharp is not your advocate.

*(because I don’t really have much of it- all but one BCS formula has been released to the public. I know enough to understand these computer formulas are more consistent and objective than any human polls, simply due to the fact that computers, unlike coaches, are not in a conflict of interest.)