Sunday, September 30, 2012

The Athlon Sports College Basketball 2012-2013 Ivy League Preview

Above, the Athlon Sports College Basketball 2012-2013 Ivy League Preview.


Anonymous said...

I'm glad to see that the preseason guides have started to publish their predictions. Shortly after Casey and Curry withdrew from Harvard, you tweeted that their absence for the 2013 season would cheapen or taint the championship if somebody other than the Crimson won.

I disagree. The unfortunate absence of two of the league's best players is an exogenous variable, similar to an injury or a player being ruled academically ineligible. Sure, that makes it easier for another team to defeat the affected squad, but "them's the breaks."

The Cornell Basketball Blog said...

Did not say it would cheapen the title. But said that critics will place a footnote on the title. That is the reality.

Anonymous said...

One could place a footnote by many titles in the record book. This coming year, not only will Harvard be missing its two beat players, Columbia will also be without its greatest offensive threat for reasons known only to him. Preseason favorite Princeton will not have a back-up point guard for the season due to injury as it tries to replace Davis.

Last year, Harvard stayed healthy while some of its top competitors were dinged up off and on.

Even our three-year run from 2007-10 benefited from an extraordinary exogenous variable. Cornell happened to stumble upon Jeff Foote who completed our Big Three core at the exact moment when, due also to an almost unbelievable coincidence, Princeton and Penn were simultaneously weak for the first time since the Ivy League had been founded. We happened to be at our best just when Joe Scott and Glen Miller temporarily crippled the perennial Ivy powerhouses. That's just plain lucky, a fortuitous opportunity which had not occurred in the previous five decades.

The Cornell Basketball Blog said...

Agree with much of what you write.

However, observers outside the league (and Harvard fans) will continuously mention the loss of Casey and Curry whenever discussing the league standings.

As for Penn and Princeton, their best years/teams in the 2000s would not have been able to stop Cornell in 2010. That was a nationally ranked Cornell team. Penn and Princeton were never ranked in the 2000s. Cornell dominated simply because that was the best Ivy team in the last 30 years and it proved itself on the biggest stages: Kansas, NCAA Tournament etc.

Anonymous said...

I'm not a big believer in asterisks in the record book for the simple reason so many seasons could be subject to one that it renders them almost meaningless. But obviously the loss of Casey and Curry is highly relevant to who wins this season.

I agree with you that Cornell's 2010 team was better than any other Ivy squad of this century. But another website which compared all historic Ivy teams on a per-possession basis concluded that the best squads of the last 30 years were, in order: Princeton 1998, Princeton 1991, Princeton 1992, Cornell 2010, Princeton 1990, Princeton 1997 and Harvard 2012.

The Cornell Basketball Blog said...

Don't intend to open up a debate here on the issue, but Cornell's 2010 team was hands down/easily the best of the last 30 years.

Princeton '98 played a very soft schedule for a league champion and won just 1 NCAA game (against an unranked team), despite earning a 5 seed.

As for Cornell 2010:

(1) Cornell finished in the final postseason poll ranked. No other Ivy can claim this.

(2) Cornell won two NCAA games, both against Top 25 teams.

(3) Cornell's schedule was the toughest of any Ivy you mentioned, Big Red facing 3 Top 10 opponents (Kansas, Syracuse, Kentucky, in addition to Top 25 wins over Wisconsin and Temple).

The author of the so-called ranking had a very poor methodology and gave little to no weight to performance in the NCAA Tournament or on poll rankings.

Anonymous said...

Per-possession methodologies do actually include results from the NCAA tournament. They just doesn't weight those games any more heavily than the other 30-odd games in a season.

Obviously, performance in the NCAA tournament is very important in determining a team's place in history and, more pointedly, in our memories because those are high profile events under the bright lights.

But per-possession methodologies recognize that NCAA games are a small sample size compared to a season's worth of games. In any small sample, random variance will play a disproportional role. We all know that, in college basketball, one hot-shooting player behind the 3-point line can win a game almost singlehandedly. Per-possession methodologies attempt to greatly increase the sample size (by a factor of 30x or 15x for a team with one or two NCAA games, respectively) in order to control for the outsized impact of small sample variance.

The Cornell Basketball Blog said...

"Per-possession methodologies do actually include results from the NCAA tournament. They just doesn't weight those games any more heavily than the other 30-odd games in a season."

Boom. Stopped reading right there. Ridiculous. It is generally accepted throughout sports that performances in playoffs are far more important than a regular season.

While the NCAA tournament is in fact a small sample size (so is 30 games in comparison to the NBA's 80+), it is in fact what defines a college basketball season. We don't celebrate terrific regular season teams. We celebrate champions. Teams that perform when it matters most, the postseason.

The methodology described above cheapens the NCAA Tournament and the Final Four. It cheapens what matters in college basketball.

Anonymous said...

Agree w/ CBB. Isn't a great regular season team just a choker if they can't do anything in the postseason?

On a different note, what's up with the game changing recruits? Still in the works, or have they committed elsewhere?

Anonymous said...

Then how do you explain the startlingly accurate predictive ability of well done per-possession analysis?

The Cornell Basketball Blog said...

"Then how do you explain the startlingly accurate predictive ability of well done per-possession analysis?"

It isn't very predictive. Last year, the per-possession analysis predicted that Harvard would win the league by a very comfortable 3 game margin (similar to Cornell's runs of 2008 through 2010). The league ended up turning out quite different.

While the analysis did predict Harvard to finish 12-2 (clap, clap), almost any casual fan could have predicted this performance. Very few people predicted 3 losses for Harvard or a run of the table (14-0).

The analysis was way off on Penn's performance (which was the big story of the season) and painted a very different picture of the league race which actually came down to the final Tuesday.

In other words, the analysis did not tell us Harvard would win the league watching a game from their dorm rooms. Instead, the analysis predicted Harvard to cruise. And they did not.

If per-possession analysis was truly reliable, all sports quants would be millionaires after playing Vegas.

Anonymous said...

I think Anonymous 2:48 PM was being sarcastic.

I have no problem with your saying that Cornell 2010 achieved more or deserves our greatest accolades on the basis of its two NCAA wins. Winning in crunch time is indeed what it's all about.

But earlier in this thread you were making the case that the best Princeton and Penn teams would not have been able to beat Cornell 2010. My point is that your assertion is debatable simply because it's not clear that Cornell was better night in and night out.

Beating Temple and Wisconsin over the course of three days when our three-point shooters shot way over their season averages is why we love that team. They deserve their special place in history.

But it does not necessarily follow that beating Temple and Wisconsin means they would dispense with Princeton 1998 or Princeton 1991. Shooting averages regress to the mean; lucky bounces become unlucky bounces.

Let's put it this way: The most exciting Ivy game in recent memory is the Harvard-Princeton playoff game in 2011. Princeton was clutch in crunch time and fully deserved the NCAA bid. The Tigers were the better team that afternoon. But would you say that Princeton 2011 was "better" than Harvard 2011? I wouldn't. In another one-game playoff, would you bet on Princeton and Harvard? I would bet on Harvard.

In the same way, you have plenty of justification for your deification of Cornell 2010. But it's one thing to call them the greatest team in terms of what they accomplished in retrospect. It's another statement entirely to say that they would defeat any other Ivy squad of the last 30 years.

Anonymous said...

The process is a statistical analysis, dealing with probabilities, not fortune telling. Yes, Penn 'overachieved' last year.

Look at the projection for the league, top to bottom, over the last two years. Then look at the actual results.

Show me any other commentator that came remotely close to this relative accuracy.

The author will publish again prior to the start of this year. My bet is that top to bottom, he will again be radically closer than other predictions. Radically.

Anonymous said...

And how do you think Vegas sets the lines? They have their own quants.

The Cornell Basketball Blog said...

Cornell's 2010 season was not defined simply by beating two Top 25 teams, Wisconsin and Temple, in the NCAAs.

Cornell beat two BCS teams in their gyms (St. John's and Alabama). Cornell swept three A-10 teams (two on the road) (UMass, St. Joe's and La Salle).

And Cornell was 30 seconds away from knocking off the No. 1 team in the country on the road (Kansas).

Princeton's best wins in 1998 were on neutral courts in New Jersey (all at the Meadowlands, 45 min from Princeton campus) against unranked Wake Forest, Texas and NC State (none of which made the NCAA Tournament). Again 3 wins in New Jersey.

Princeton did not own any top 25 wins that season and did not play any final top 10 teams with the exception of their loss to Michigan State in the NCAAs 2nd round.

Cornell played 3 top 10 teams (Kentucky, Kansas, and Syracuse).

Princeton's final record was sexy, but lack substance. When you look at who they played, who they beat, where they won, and how they performed in the NCAAs, the season was not so impressive. Cornell's 2012-2013 schedule is tougher than Princeton's '98 schedule.

Not sure how anyone can defend Princeton '98. The whole argument is premised on the fact they lost 2 games with an easy schedule close to home.

Anonymous said...

I don't think that Anonymous 2:48 PM was being sarcastic. It's not that per-possession analysis is infallible. It's that it's the best tool AVAILABLE.

CBB can disparage Princeton '98 until he's blue in the face but the point the last couple posters made is that, in a one-game playoff on a neutral floor, Princeton '98 would be favored over Cornell '10.

That doesn't mean Princeton would win the one game. But it does mean that, until the betting public forces Las Vegas to move the line, the sportsbooks would put their own money on Princeton winning.

More importantly, it means that if Princeton '98 and Cornell '10 were to play 100 games, we would expect with a degree of confidence we could quantify a priori that Princeton would win more than 50 of those games.

That makes Princeton "better" even if Cornell is the "greatest team" of all time.

The Cornell Basketball Blog said...

Cornell had a better body of work and was the better team. Not just in the eyes of this author by in the eyes of the national media.

The Princeton '97-'98 team did not receive much attention (even back then)... because frankly, the schedule was soft and they did not win when it mattered.

Princeton '98 was the "could shoulda" team. Cornell '10 was the "did that" team.

Anonymous said...

Other than being ranked eighth in the country, the Princeton '97-'98 team did not receive much attention.

The Cornell Basketball Blog said...

Funny how you say they "finished" ranked 8th. Actually, they did not finish ranked 8th. You cite the last poll BEFORE the Tournament, before they lost to Michigan State. Ranking fell off after the 2nd round loss. Cornell however finished ranked #17.

Similarly, Penn was in the Top 25 as well in 1994, but fell out BEFORE the final poll. And Harvard was in the Top 25 in 2012, but fell out BEFORE the final poll.

If you think Princeton '98 received tons of media attention, feel free to start pasting links here to all of their media attention. You will not find much. There was very little ink on them considering they reached the top 10.

The lone "negative" on Cornell 2012 resume is the loss on the road to Penn. Aside from that defeat, Cornell was just about perfect.

Have yet to see one comment in this thread substantively support Princeton as the better team. Who did they beat? What wins were so impressive? Why would anyone celebrate 3 BCS wins in the Meadowlands as the best team ever? Princeton '98 did not have a single win that was better than Cornell's wins at Alabama and at St. John's. And those were not even Cornell's best two wins!

The Cornell Basketball Blog said...

Princeton's best wins in 1998 came over two NIT teams, Wake Forest and NC State. The third best win was over Texas. Longhorns did not even make the postseason. Of these three wins, each were in New Jersey in the Meadowlands, 45 minutes from Princeton campus.

Before we even get to Cornell 2010's best two wins on a neutral court over two Top 25 teams (Wisconsin and Temple), Cornell owned two true road wins at St. John's and at Alabama. St. John's was a 2010 NIT team. These two road wins are arguably superior to Princeton's three neutral court wins (which includes the two NIT teams).

Princeton did not own any wins over any team in the 1998 NCAA Tournament. Cornell had a win over Vermont, a 2010 NCAA Tournament team.

Cornell 2010's 5 losses were to three final Top 10 teams (Kentucky, Kansas and Syracuse) plus Penn and Seton Hall (NIT team).

Princeton had just two losses, both very respectable, North Carolina and Michigan State.

So Princeton had no bad losses, but Princeton also had no real high quality wins-- just two neutral court wins over a pair of NIT teams.

To even suggest Princeton '98 as the best ever is to completely ignore their resume.