Why the ACC is Going to Win the ACC/Big Ten Challenge

*note: this is part 1 of the ACC/Big Ten Challenge Preview. Check back later to read Andy Bottoms' nonsense about how the Big Ten will win. As if.

The 13^th annual ACC/Big Ten Challenge begins tomorrow, and prospects for the ACC are grim. According to Ken Pomeroy’s site, the Big Ten is the strongest conference in the country, and it’s not even close. The ACC ranks a distant 5th (behind the SEC! Seriously, Ken?). The Big Ten has won two Challenges in a row. They are favored in 9 of the 12 games. The bottom end of the ACC is terrible. Boston College has 9 freshmen and looks to make last year’s Wake Forest team – which was historically bad – look like a fairly solid team. Boston College’s two victories this season were by three over New Hampshire (who’s only win is over an NAIA team) and in overtime against UC Riverside. Wake, meanwhile, is one of three ACC teams giving up more than a point per possession to suspect competition. Virginia’s offense has been woeful. Florida State tied an NCAA record for fewest combined points in a half (a 14-14 barnburner against Harvard). Miami is without Reggie Johnson – the best big man in the ACC. Clemson lost back to back games against the College of Charleston and Coastal Carolina. Georgia Tech lost to LSU, which is one of the worst major conference teams in the nation. Maryland lost their point guard and is left with a roster so thin that a walk-on is playing 10 minutes a game. And North Carolina’s Harrison Barnes left the UNLV loss on crutches. So will the ACC win the ACC/Big Ten Challenge? Of course not. Hell yeah!

The reason the ACC will win is because of proportional win-shares and the bell curve. I was first introduced to proportional win shares at Bud Elliot’s excellent blog Tomahawk Nation. You can read his entire treatise here, or bear with me and I’ll explain it the best I can.

Say you predict all 12 games of the ACC/Big Ten Challenge, and to do so you go with the favorites. According to Ken Pomeroy’s site that list would look like this:

Add that up and the Big Ten wins 9-3. Seems reasonable, right? Teams will take care of business, or – if there are upsets - the number of upsets will be balanced. UNC might beat Wisconsin, but then Northwestern would knock off Georgia Tech – or some such scenario. If I were to offer you a bet that you win with 9 or more Big Ten victories and I win with 7 or fewer – with 8 being a wash -  does that sound like decent odds? I mean, hold serve or have an even number of upsets and you win. With one upset in my favor it’s still a wash. If you answered yes, you’re not a gambler. Put your money in your pocket. That’s not how gamblers look at games, and you shouldn’t look at them that way either. So here’s that same table, only with a significant tweak. Instead of a predicted winner, that column is now replaced with the statistical probability that the Big Ten team will win. These percentages are based off the point spread. A pick ‘em is a 50/50 game, a one-point spread is a about a 54% probability, etc...all the way down to a 1-seed having greater than a 99% chance over most 16-seeds.

Alright, now we’re getting somewhere. This final table is just Table 2 repeated, except that the percentages have been replaced with their decimal equivalent. And those decimal numbers are what I was referring to by a proportional win share. For example, a team with a 60% probability of winning would win – on average – six of every ten games played (and lose four). So divide six (the number of likely wins) by ten (the number of chances) and you get 0.6, which is a proportional win share.

Now, for the fun part which will make you feel like a savvy gambler, sum the total of those numbers. Or not, and I’ll do it for you: 7.28. And that – according to Ken Pomeroy – is the predicted number of wins for the Big Ten. So even though they are favored in 9 of the games, the odds are that they’ll finish with 7 wins. Math!

Of course that still leaves the victory in the hands of the Big Ten. And this is where the bell curve comes in. Without getting too geeky a bell curve is just the normal distribution of a data set. For example (assuming Pomeroy’s odds are right), if the 2011 ACC/Big Ten Challenge were played 1,000 times then the most common outcome would be the Big Ten winning 7-5. 8-4 would be the next most common, followed by 6-6, etc... And there would be times – maybe not many – where the ACC would not only win, but win by as much as 8-4 or even (gasp) 9-3. The odds aren’t great, but they’re there, and that’s why the ACC is going to win. This particular version of the 2011 ACC/Big Ten Challenge is going to be an outlier on the edges of the bell curve.

Here are the games, with Pomeroy’s Big Ten probabilities, and why he’s wrong.

Miami (4-1) at Purdue (6-1) (Purdue 79%)

Miami might have the best guard tandem in the ACC in Malcolm Grant and Durand Scott. But Durand Scott is off to a very inefficient start. Grant does damage from the outside, whereas Scott is the one who can do it inside the arc. Last year Scott converted 46% of his 2s and 39% of his 3s, whereas this year those same numbers are 31% and 33%. Time for those to change.

Should be Purdue 65%

Georgia Tech (4-2) vs Northwestern (5-0) (Northwestern 48%)

Sure, Northwestern is 5-7 in past challenges and Bill Carmody’s club appears to be legit. But they’re Northwestern! Pomeroy has to be wrong.

Should be Northwestern 40%.

Maryland (3-2) vs Illinois (6-0) (Illinois 55%)

This is a tricky one because last year’s Maryland team was overvalued by Ken Pomeroy, and I’m guessing this year’s team is as well. It’s a bit less than a coin flip, which means that Illinois would be a significant favorite if at home. The difference here is going to be free throws. Maryland is shooting 59.3% while Illinois is shooting 77%. Maryland isn’t really that bad, and Illinois isn’t really that good. Time to move toward the mean.

Should be Illinois 42%.

Virginia (5-1) vs Michigan (5-1) (Michigan 38%)

Michigan is giving up 34.1% from deep (180th in the country). Virginia, meanwhile, is shooting 32.5% (197th). The problem is that last year Virginia shot 38.8% (19th) and has virtually the same shooters who were making them last year missing them this year. Gotta change sometime. Might as well be against Michigan.

Should be Michigan 20%.

Clemson (3-2) at Iowa (4-2) (Iowa 45%)

Who has had the fewest of their shots blocked this season? Clemson. Who’s 316th iin blocked shots? Iowa. Time for 5* bust Milton Jennings to shed the ‘bust’ from his title. Go to work, big fella. Do it for the ACC.

Should be Iowa 33%.

Duke (7-0) at Ohio State (6-0) (Ohio State 78%)

Another game where a return to the mean on free throws will help the ACC. Only 2 of 345 schools have suffered through watching their opponents make as high a FT% as Duke (81.5%!).  Ohio State is about 1% below the NCAA average, and that’s with their bell-cow Jared Sullinger shooting over 85%. Last year he was a 70% free throw shooter, and even assuming he’s improved he shouldn’t finish the season with more than 75%.

Should be Ohio State 55%.

North Carolina State (5-1) vs Indiana (6-0) (Indiana 53%)

Indiana is Indiana but NC State has the most Indiana player of all. Scott Wood is from Marion, Indiana where I’m sure he honed his jumper against the side of a barn when he wasn't too busy de-tassling corn. Marion was the birthplace of James Dean and Cole Porter, and like those legends Scott Wood – in the words of Big 10 Network announcer Gus Johnson – is cold blooded. Last year he was the best long range shooter in the ACC. In Wood’s first ACC/Big Ten Challenge he went 0-6 from deep. Last year he went 3-6. This year he’s due for 6-6.

Indiana, literally, has no chance.

Boston College (2-4) vs Penn State (5-2) (Penn State 62%)

Boston College lost to Holy Cross by 22 and UMASS by 36. If Ken Pomeroy says that they have a 38% chance of beating any team with a pulse then I’ll take it and back away slowly.

Should be Penn State 62%.

Florida State (5-2) at Michigan State (4-2) (Michigan State 68%)

Michigan State has an elite defense combined with the 9th ranked offense in the Big Ten. Florida State has an eliter defense combined with the 9th ranked offense in the ACC. This is a game children shouldn’t watch. FSU is 302nd in turnovers. Michigan State is 262nd. There can’t be a line on a game like this.

Should be Michigan State 50%.

Virginia Tech (5-1) at Minnesota (6-0) (Minnesota 62%)

Tubby Smith is the coach of Minnesota, and Tubby Smith is 1-3 in the ACC/Big Ten Challenge. His lone victory came over a bad Virginia team in 2008. That same Virginia team beat Virginia Tech late in the season and cost the Hokies a bid to the NCAA Tournament. And Seth Greenberg is still pissed.

Should be Minnesota 55%.

Wake Forest (4-2) at Nebraska (4-1) (Nebraska 86%)

Last year Wake Forest was the worst team in the history of the ACC. And they still won in the ACC/Big Ten Challenge. This year they’re taking care of the ball (40th nationally, compared to 272nd) and Nebraska isn’t forcing takeaways (281st). That will mean plenty of extra possessions for Wake to throw something toward the rim that will be possibly go in.

Should be Nebraska 75%.

North Carolina (5-1) vs Wisconsin (6-0) (Wisconsin 54%)

Wisconsin is 344th in the nation at getting to the line. UNC is the third best team in the nation at keeping teams off the line. Wisconsin may not shoot a free throw all night, while UNC averages 28 attempts a game, and has shot a minimum of 11 more than their opponent in every game. This means that while UNC is parading to the line Twitter will be erupting with tweets about how the Badgers are getting ripped by the refs. And nothing is more annoying than the combination of incorrect ref bashing and tweets.

Should be Wisconsin 50%.

Here’s are final look at the table, with the proper proportional win shares according to where this Challenge will end up on the bell curve.

Sum the right hand column: the Big Ten has 5.47 wins coming their way. And that is why the ACC is going to win the ACC/Big Ten Challenge.