Requesting Statistical Assistance

by jtothep on Sep 2, 2010 3:39 PM EDT

Hi, apologies in advance if this makes me look more ignorant than I've already come across in here, but I'm still trippin on a percentages puzzle posted by our friend PSUinBosston in last night's thread. The assertion: If each year we play 3 teams from the other division (inter-divisional game) and we go on to win our division, Then we have a 50% chance of meeting one of them in a CCG rematch. On the surface, it seemed sound enough to shut me up, but I think I need the input from Ibleedbluenwhite or one of BSD's other statisticians, b/c it doesn't feel right. /venus'd

I think we learned late in the broadcast last night that results from inter-divisional games count toward Division champion standings. They didn't specify how exactly (if intra-divisional games are tallied up first, then you move to inter-game results as a tie-break), but would it even matter? Doesn't the chance of rematch actually dwindle the moment the result is posted, because one of us then has a blemish on our record? Anybody have some good math insights? Thanks.

Thanks for the fanpost! Please do not post any content from a premium site that requires a subscription. Also, if you just want to share a link, quote, or video, please consider using fanshots instead. Thanks.

Tweet Comment 22 comments | 0 recs |

Do you like this story?

Comments

I looked into this

I was going to try to refute the UM and OSU claims about meeting twice in a row by showing the odds of a rematch were slim. I thought like you did: once a game is played and one team is credited with a loss, the chances of winning their division drop.

I looked back over the past ten years at the SEC, ACC, and Big XII, and surprisingly 48% of the time (12/25) the CCG was a rematch of a regular season game. I think in 10% of those matchups the team that won the first game lost the second game.

Mike
@BSDtweet
Black Shoe Diaries

by BSD on Sep 2, 2010 3:52 PM EDT reply actions

That's really quite shocking

Any chance your data is pastable? In image format maybe?

And they all three do what we’ve got lined up, right? 5 games inside and 3 games outside? But do they keep separate Division Standings (as opposed to Conference and Overall standings)?

Until our defense proves otherwise, it should be presumed they will be excellent.

by jtothep on Sep 2, 2010 4:13 PM EDT up reply actions

Not sure if I saved it

I’ll have to check my computer at home. The spreadsheet was minimized on my desktop for days before I closed it though. When the stats didn’t come out the way I expected them to I canned the post I was working on.

Mike
@BSDtweet
Black Shoe Diaries

by BSD on Sep 2, 2010 5:12 PM EDT up reply actions

Interestingly

I didn’t examine this closely, but if you look back I think you would find that usually when you have a rematch it is because you have one division that is really weak where the winner is maybe 3-2 in the division and wins a tiebreaker of some sort. Either that or there is one fish in the big pond type deal. Rematches in the SEC are rare (I think the last one was in 2004) because the competition is so good that if you drop a game in conference you probably aren’t winning your division. I think we’ll see the Big Ten operate along these lines too.

I would have to check the schedule, but I think the B10 really loaded up the 2011 schedule so all the heavyweights are playing each other. So it wouldn’t surprise me if we get a rematch. It may also open the door for someone like Northwestern to sneak into the CCG.

Mike
@BSDtweet
Black Shoe Diaries

by BSD on Sep 2, 2010 5:18 PM EDT up reply actions

an image of data?

Who do you think we are GT?

BTW: never do a google image search for “chart” at work…you’ll never get anything done.
see!!!

"I have my Joe Paterno autograph already, but I don’t know that I’d begrudge anybody else from getting theirs no matter their age. That’s kind of like meeting Winston Churchill." jesse. @ BSD

by bconway6 on Sep 2, 2010 6:07 PM EDT up reply actions

pretty!

Joe Paterno - Not Unlike Hugh Hefner

by letsgopsu on Sep 2, 2010 6:13 PM EDT up reply actions

I'll see your Google images search and raise you one EDSBS.

Boy do I ever love their charts!

by coachKofPA on Sep 3, 2010 6:28 AM EDT up reply actions

Wow, talk about a fail for UNC.

If only that was the case for their men’s basketball team…

Oh no, don't tell me Matt Barkley ALSO doesn't tie his shoes...
September 1st, 2010. A day that shall live in infamy.
"Because one of the great minds of the 21st century is raising glow-in-the-dark fish and weaving serapes..." -Leonard Hofstadter from The Big Bang Theory

by OBrienSchofieldismyHero on Sep 3, 2010 2:05 PM EDT up reply actions

Did you sleep through the last basketball season?

by Frank O'Brien on Sep 3, 2010 2:41 PM EDT up reply actions

No, I meant during the Hansborough Ascendancy.

That team was sickening. So much help from the officials…

by OBrienSchofieldismyHero on Sep 3, 2010 3:14 PM EDT up reply actions

Would you also factor in probability of winning the division must on past performance and/or perceived strength

Odds would be slightly higher for a PSU/Nebraska rematch in the CCG vs. say MSU and Indiana.

by Frank O'Brien on Sep 2, 2010 4:02 PM EDT reply actions

Would they, though?

Because remember that PSU/Nebraska are automatically playing a tougher schedule, as are OSU/UM (usually) because they’re playing good teams OOD (out of division). Right?

"Every player we have, someone-maybe a parent, a grandparent, someone-poured their soul into that young man. They are handing that young man off to us. They are giving us their treasure, and it's our job to make sure we give them back that young man intact and ready to face the world."

-J.V.Pa.

by psume06 on Sep 2, 2010 4:05 PM EDT up reply actions

In other words, I think its a wash - MSU is getting an easy win, while the winner of Neb./PSU is getting a tough win.

But they still have to go through the other top teams…?

-J.V.Pa.

by psume06 on Sep 2, 2010 4:05 PM EDT up reply actions

We're talking specifically here about a rematch in the CCG

Bottom feeders like Indiana and Michigan (not MSU in this case) will have a far smaller chance of reaching the Championship Game. You are on to something here though that should not be overlooked, the big name programs, Ohio State, Nebraska, Michigan and Penn State have a decidedly tougher schedule year in and year out with the protected rivalry game. Except of course tOSU getting their cupcake at the end of every regular season.

by Frank O'Brien on Sep 2, 2010 4:25 PM EDT up reply actions

No

And here’s why I need bleedbluenwhite. I know it’s a pure (no qualitative elements included) numbers game—quant only.

I’m thinking spatially, in terms of slots, but there are too many scenarios. 8-0 wins your division every year, but that same year 4-4 might win the other one.

Until our defense proves otherwise, it should be presumed they will be excellent.

by jtothep on Sep 2, 2010 4:25 PM EDT up reply actions

Wow

You thought about this much more deeply than I. Why do I have a feeling I am going to come out looking like a total ass on my 50%?

by PSUinBOSSton on Sep 2, 2010 4:57 PM EDT reply actions

47.5% isn't gonna make you look stoopid

And as far as I can think, the only “factor” that plays into this thing not being a dead 50% is that you’re probably presuming that the Division X team beats all of its rivals in Division O… These assumptions could create an issue, but I wouldn’t think much.

-J.V.Pa.

by psume06 on Sep 2, 2010 5:42 PM EDT up reply actions

Youre right if all things are random...

Given: Team A is in the CCG.
Given: Team A played 3 of the other 6 teams of the other division.
Given: Team A plays a random team from the other division. (this is true if each team from the other division is equally as likely to win their division. In practice, team strengths and protected games affect the randomness.)

3 teams out of 6 is .5 = 50%

I’m a little surprised that the only reason for the 47.5% is that an odd number of games had been played.

"I have my Joe Paterno autograph already, but I don’t know that I’d begrudge anybody else from getting theirs no matter their age. That’s kind of like meeting Winston Churchill." jesse. @ BSD

by bconway6 on Sep 2, 2010 6:12 PM EDT up reply actions

as I was starting to state in a post that I accidentally just deleted

but basically it is what the bacon way says, except that again, once you beat that team from the other division, they are now a bit less likely to win their division. To balance that out, though, is that so would (theoretically) the other 2 teams in that division that you would have likely beaten. So now there’s a slight advantage for the 3 teams you did NOT beat, but more than likely they also have a loss, so then it really comes down to how good all the individual teams are.

If you played and beat the best team in the other division, and you are their only loss, meanwhile they beat everyone in their division, then all those other teams would have at least one loss as well. I think one of the tiebreakers is divisional records, so that would actually increase the chance of replaying that team.

Things get even harder to figure out when you start talking about multiple loss teams, but I doubt the percentage would drop to that much less than 50% anyway.

by The JuggerNitt on Sep 2, 2010 6:19 PM EDT up reply actions