Friday, August 9, 2013

News and Notes: Friday Edition

Below, news and notes for Friday...

  • The Cornell Rebounder Club's  "Welcome Back Picnic" is just around the corner on August 28.  If we are all lucky, Cornell Athletics will give us some video footage as they have in the past.
  • Princeton released its schedule and previewed Cornell, noting, "First-team All-Ivy Leaguer 6-7 Shonn Miller (11.5 ppg, 6.8 rpg) is back, though Cornell's next two leading scorers, 6-3 Johnathan Gray (10.0 ppg) and 6-6 Errick Peck (9.7 ppg) are not. Twelve Big Red players got a start last season, and five of those players are on the upcoming season's roster."
  • Below is a tentative listing of opponents on Cornell's 28-game 2013-2014 schedule which will include 6 home nonconference games.  The games listed are in no particular order. Teams which participated in the 2013 postseason are noted.  Cornell's nonconference schedule will likely feature three Top 35 preseason teams, two in the Top 10 (Syracuse is one of them).  The opponents:
  1. at Syracuse (2013 NCAA Tournament Final Four) (Season opener)
  2. vs. Loyola (MD) (2013 Invitational Tournament
  3. vs. Binghamton
  4. at Western Michigan (2013 College Basketball Invitational)
  5. at St. Bonaventure
  6. at Stony Brook (2013 Postseason National Invitational Tournament)
  7. at St. Peter's
  8. at Colgate
  9. NOVEMBER 22, 7 pm, vs. Siena 
  10. NOVEMBER 25, 7  pm vs. Radford
  11. DECEMBER 7, vs. St. Francis (PA)
  12. at BCS Opponent (Contract signed, "NCAA Tournament team")
  13. at BCS Opponent (Contract signed, "NCAA Tournament team")
  14. vs. Division III Opponent
  15. PLUS 14 Ivy League Games


Anonymous said...

Getting mauled by Syracuse is a rough way to start the year.

The Cornell Basketball Blog said...

Probably not even Cornell's toughest game.

Anonymous said...

What's the point of playing these programs when we're so clearly outgunned?

As an aside, your twitter post comparing Harvard's upcoming softish schedule to the 98 Princeton team is a bit of a stretch: that Princeton team played a ranked Texas team, a UNC team that was ranked #1 in the country and had NBA All Stars on it, and other games against Wake Forest, NC State, Drexel, etc. Harvard isn't playing anyone close to being a contender to be national champ.

The Cornell Basketball Blog said...

The ONLY team Princeton played on its conference schedule that made the NCAA Tournament in 1998 was North Carolina and Princeton lost that game.

None of the other teams made the NCAAs. The fact that some of those teams were ranked in the preseason and in November is irrelevant. They were bad teams that did not make postseason.

Cornell played two teams in the regular season of 2010 that FINISHED the year in the Top 10 (Kansas and Syracuse).

Princeton '98 and Harvard '13-'14 = soft schedules.

Anonymous said...

You seem to have some indication as to who the other D1 opponents are. Can you at least give us hints?

The Cornell Basketball Blog said...

The only D-Is that you don't know about are the two other BCS opponents joining Syracuse on the schedule.

One is a lock for Top 25. The other could be preseason Top 25 and a lock to at least get votes.

All three BCS teams are household names, big time basketball programs with winning traditions in basketball.

Fans should be happy with the two other BCS games.

Anonymous said...

With regard to your obsession with the 1998 Princeton team, Billy Shakespeare said it best, "Methinks the lady doth protest too much."

The Tigers led Antawn Jamison, Vince Carter and the rest of the #1Tar Heels deep into the second half, at the Dean Dome no less. You dismiss that game as merely a loss, the same way your beloved RPI treats a one-point overtime loss the same as a 30-point blow-out.

The tournament committee thought enough of that Princeton team and their schedule to see the Tigers fifth, seven lines higher than our 2010 team.

Admit it, the main reason you are fixated on the 1998 Princeton team is that Mike James' model says they are the best Ivy team of the AI era, and that grates at you. Princeton could lose an overtime game to an NBA All-Star team but you would still point out that it was a loss.

BigRed1965 said...

The pictured program (which I have in my collection) brings back memories. That NYU team was slated to be top #1 or #2 in the country, led by Barry Kramer and Happy Hairston. Cornell actually took a 32-32 lead before the superior power of NYU prevailed for a final 82-65 score. Hopefully our current team will do better in its opener against Syracuse, but it has been 45 years since we've beaten the Orange, hasn't it?

The Cornell Basketball Blog said...

And Cornell 2010 was down by 6 with 5 minutes to go against Kentucky, the FIRST TEAM IN HISTORY to have five (5) 1st round NBA Draft picks. And... Cornell was down by a bucket at the nation's longest home court winning streak at #1 Kansas with under 3 minutes in the game.

Princeton beat nobody. Again, they don't own a single win worth bragging about.

12 seed UNLV was the best team they beat all season long. The only team they beat to make the field of 64.

They had a couple BCS wins in NEW JERSEY on a neutral court against non-qualifying teams.

Cornell won ON THE ROAD... at Alabama, at St. John's, at UMass, at La Salle, had TWO Top 25 wins over Temple and Wisconsin and beat NCAA Tournament team Vermont.

You cite Princeton's defeat to UNC. Well, I just doubled you up with good losses vs. Kansas and Kentucky.

Cornell went 13-1 in the toughest Ivy to date, won games by an average of 17 points per game against teams that included Princeton (CBI semifinals) and Jeremy Lin's postseason Harvard team.

Princeton in 1998 beat a bunch of non-NCAA qualifiers on a "neutral" court, 45 minutes from campus in East Rutherford. Then Princeton dominated an Ivy League that was purely terrible aside from a mediocre Penn team.

Mike James' model is flawed because it relies too heavily on point differentials and not enough weight on actual high quality wins, especially road wins. Road wins and Top 25 wins are what show the character of powerhouse teams.

Winning by 30 over Brown is not high quality evidence of being a national power.

To claim you are the best, you need to be the best.

Cornell 2010 won just too many tough games on the road and vs. Top 25 teams. Then Cornell had the respectable losses against Kansas and Kentucky.

There is just not enough on Princeton's resume to compare.

If you want to argue you "think" Princeton would win, well, you are entitled to your opinion.

Having watched both teams, I think Cornell would have won. (That Princeton team was very similar to the Wisconsin team we beat)

But regardless of your view on which team would win, there is absolutely no question which team put together a better resume.

Princeton had a 5 seed because they entered the postseason with just 1 defeat at North Carolina.

Where do you think the Committee could place them? You can't place a 1 defeat team on the 12 seed line.

If anything, the 5 seed was an INSULT. They were the #8 team in the country and they were relegated to a 5 seed with fringe top 25 teams.

They took care of business against a 12 season (UNLV) then were dispatched from the tournament by a 4 seed.


Anonymous said...

I wouldn't be so dismissive of "mediocre Penn teams." In 2010, Cornell lost to a Penn team which was worse than mediocre.

Anonymous said...

I am actually kind of surprised that so many people insist on letting a MODEL decide who is best. Mr James keeps going on and on about Cornell's early-season narrow wins over bad teams, like Bucknell, and late loss to Penn, as if that discredited the number of great teams we played close (Kansas, who were #1, so I don't know if we were supposed to be floored by the fact that Princeton 98 played UNC close for three quarters of a game or whatever), or the fact that they actually BEAT ranked teams and progressed farther in the tournament.

I figure that it was obvious the Penn loss was a fluke, but apparently no, throw it into the "model".

Northern Iowa, who defeated ranked teams that year and ended up booting Kansas, lost to the two WORST teams in its conference. Did any rational individual think it made them not a great team? No. Everyone just noted that the best teams in a conference sort of get complacent sometimes and then have random losses. But a guy like Michael James will just dump it all into his "model" and say it makes a team worse than another team who couldn't play ranked teams like they could.

Honestly? I used to find Mr. James pretty rational until he started letting his "model" define greatness.

Anonymous said...

Also, the Blog is inadvertently short-changing the 2010 team. We were actually up one at Kansas with under a minute to go, as Foote dropped in a go-ahead bucket for our last lead that night....

Anonymous said...

Also, the Blog is inadvertently short-changing the 2010 team. We were actually up one at Kansas with under a minute to go, as Foote dropped in a go-ahead bucket for our last lead that night....

Michael James said...

Hopefully this response doesn't also get eaten by the Blogger monster, but here's my second crack at it.

A) I don't want to take credit for any model being mine. I did not create the model I use, though I have made adjustments to it to try to improve its predictiveness. The fundamental concept for the model was dreamed up by folks way better at this than I.

B) "My" model is not the only one that believes that Princeton 1998 was better against 1998 competition than Cornell 2010 was against 2010 competition (I make that point, because I don't want to opine as to whether 1998 college basketball was better or worse than 2010 college basketball).

C) Other "Adjusted Scoring Margin" models that believe Princeton 1998 was better than Cornell 2010 include College Basketball Reference's SRS model (, and the Massey Ratings (,

D) It's not just the Adjusted Scoring Margin models that rate Princeton higher. It's the win-loss only models (the RPI) as well (, NOTE: The 1998 one wasn't final, but I assure you that Princeton didn't fall 16 spots for going 1-1 in the NCAA Tourney.

E) It's also the goofy human polls that thought Princeton was better (#16 vs. #17 in the coaches).

Some of these differences are marginal, while others are more pronounced, but the common thread is that they all point toward Princeton 1998.

Cornell obviously had the best March run of any Ivy in the AI era. The Big Red is clearly number one there. On its best days, Cornell 2010 was simply amazing with seven of the Top 200 individual Ivy game ratings of the AI era (over 7000 games). But the 1998 Tigers had nine such games.

Princeton 1998 also only had two games in which it played like a below average team nationally (adjusted pythag win% below .500). Cornell had five.

The point of ASM models is to understand what a team is like on its average day, and how much variance there is in that average performance. Cornell 2010 was much higher variance than Princeton 1998 with only a slightly lower average, so the Big Red it logically follows that the Big Red would see higher highs and lower lows. Pointing to its highest high as representative of its likely output is as silly as pointing to its lowest low. That isn't the same question as "If I had to pick one Ivy to beat the best team in the nation, which would I pick?" in which case, I'd more heavily favor a higher variance squad, even if it had a slightly lower average expectation.

The question I am trying to answer is "Which Ivy team was the best over the course of the season in which it played?" The preponderance of evidence points to that being 1998 Princeton. If you asked me which team I'd want to play one game against the best college team in the nation... my answer might very well be different.

The Cornell Basketball Blog said...

The "goofy human polls" (i.e. the professional/coaches who get paid to run programs) did not compare Princeton '98 against Cornell '10.

Princeton was ranked for its particular season, not against other seasons. I should also mention, I have yet to see the 1998 poll on a reliable website. But for the sake of this argument, we will assume the poll is accurate.

But again, being the 16th best in 1998 does not even guarantee they were a top 25 caliber team for 2010.

Further, the numerical models you cite are all flawed.

At the end of the day, Princeton did not prove it could beat teams at the higher ceiling levels (the top 35 caliber teams) or other high quality teams on the road.

And frankly, I am not going to favor any team that does not have high quality wins and a strong road record.

But I understand that Princeton was 27-2, a sexy 27-0 record against forgettable teams (and virtually no road challenges).

Anonymous said...


Thank you for your detailed explanation. That was informative. Would you say that, in general, the greatest determinant of whether a team displays high variance is the extent to which it relies upon three-point shooting?

CBB believes that, because Cornell was "on" from the perimeter against Temple and Wisconsin, this means the 2010 Big Red team would therefore be "on" against every team that it ever played. You know, because if you flipped a coin twice and it came up heads both times, it will always come up heads.

The Cornell Basketball Blog said...

At the end end of the day, Princeton '98 did not (and cannot) prove it was capable of winning a top 25 game or a quality game on the road because the Tigers never did it.

Woulda, coulda, shoulda.

Princeton beat average to below average teams and did it by large margins. That's their resume.

Taking those scoring margins over average teams in Rutherford, NJ, you can crunch the model all you want. You can argue projections, but again, at the end of the day, the real substance is not there. Princeton did not win the games it needed to win to make the argument Mike wants to make.

Mike's argument relies entirely on what he thinks Princeton was capable of (against elite competition) but not what they actually did.

At least Cornell can say, they beat two Top 25 teams. At least Cornell can say they went into the arenas of athletic/BCS teams on the road and took wins.

Princeton can't do either. But Princeton can brag that it won a bunch of Ivy games by 25 points.

Michael James said...

Three-point shooting is a massive determinant of variance, not just on the offensive end, but on the defensive end as well. If you look at the five highest variance teams of the past 10 years on each side of the ball, all of them on the offensive end (PRIN 06, COL 07, BRN 04, COR 10, BRN 09) either shot a Top 100 ratio of threes or derived a Top 100 percentage of points from threes (and usually both).

On the other side, there are three Sydney Johnson PRIN teams among the lowest variance defensive sides, because he's never coached a team that lets opponents shoot threes (he's been in the Top 50 in fewest 3pters allowed all six seasons at the helm of Princeton and Fairfield).

There are other elements of the game that affect variance as well. The free throw line is a source of steady points, and if you don't go that often, you are forced to score with actual shots. Teams like Brown 2004 and Princeton 2006 got to the line at rates among the lowest nationally, which when combined with their high 3PT rates, means that if the 3s aren't falling, there aren't a lot of steady points to fall back on. In a similar vein, on the defensive end, if you're sending opponents to the line all the time, you'll probably have lower defensive variance, because your opponents are starting off every game with a consistent number of free points.

Pace is another factor, though that requires real effort to deviate far enough to affect variance (like the difference between Glen Miller's Penn teams and Joe Scott's Princeton teams).

As for the argument about who each team played... Here are the games each team played against opponents with an SRS score above 5. I've blinded the opponents, I'm just going to show you SRS score (higher the better, read it as the expected margin on a neutral floor over a national average team), location, result and MOV.

SRS - 25, Away, Loss, -5
SRS - 22, Away, Loss, -15
SRS - 21, Neutral, Loss, -17
SRS - 18, Neutral, Win, +18
SRS - 12, Neutral, Win, +13
SRS - 10, Home, Loss, -10
SRS - 10, Away, Win, +4
SRS - 9, Away, Win, +5

SRS - 27, Away, Loss, -8
SRS - 19, Neutral, Loss, -7
SRS - 11, Neutral, Win, +2
SRS - 10, Neutral, Win, +5
SRS - 9, Away, Win, +12
SRS - 7, Neutral, Win, +12
SRS - 5, Neutral, Win, +6

Now, this ignores the fact that Cornell either lost or won by single digits against 10 teams not listed in that above sample while Princeton won just three by fewer than that mark. But it's important to show because college basketball teams are measured on a gradient of skill, not measured by a binary "in the Top 25/not in the Top 25" metric. It's why so many Top 25 teams are underdogs when playing on the road against Top 50 teams in conference play. The difference between the two squads isn't even enough to cover the HCA.

The Cornell Basketball Blog said...

I think everyone understands variance. But the heavy reliance on scoring margins (and lack of variance) causes a model to miss the mark.

"Best Ever" teams are universally defined by (1) wins against the very best/top competition, (2) resilience (i.e. top road wins), (3) wins that matter most (i.e. playoffs).

We as a society do no celebrate big wins in the preseason and beating up on cupcakes.

Princeton '98 did not have #1 or #2 above. They are unproven. (They did have an important playoff win vs UNLV somewhat satisfying #3).

Any model used to measure "best ever" --- should and must place significant weight on Top 25-40 wins and road wins against top competition.

You still can't point to a high quality win by Princeton (Top 25 or on the road). Calling them "the best" based on merely the stats they pumped up on mediocre competition is silly.

But yes, Princeton '98 was very consistent against average competition and proved it could win in New Jersey against teams not good enough to make the postseason. Congrats.

Anonymous said...

just stop.

mrjames is a stats nut.

You won't be able to use any sort of reasoning if it isn't stats backed (and backed by his models).

Moneyball has made people think that stats are the end-all, be-all, but in reality there is much truth to the blog's stance.

I'd rather have a team average 1 point wins against top 25's AND the ivy league, then a team unproven against top 25's and win by 20+ points against the ivy league. One team closes games and the other beats up on a crappy league.

The Cornell Basketball Blog said...

I understand Mike James' argument. But his argument of Princeton '98/"best ever" lacks any real substance when you look past the influence of the average margins of victory.

Again... look at Princeton '98's resume (or lack thereof):

ZERO Top 25 wins.

ZERO wins during regular season against teams making the NCAAs.

ZERO road wins against BCS teams outside of New Jersey.

And then Princeton '98 went 14-0 against a very poor Ivy League conference slate.

Princeton's best win of the year, by far, was the 1st round game against unranked, 12 seed UNLV.

Princeton's only road win arguably of note (teams from BCS/A-10) was Rutgers (14-15 overall)

Princeton's other best overall wins were in New Jersey in the Meadowlands (alleged "neutral court") against non-NCAA Tournament teams (NC State, Wake Forest, Texas) by a combined 13 points.

That is it. Princeton's resume boils down to UNLV nd 4 BCS wins in New Jersey (all non-NCAA Tourn. teams). That is Mike's best case on its best day.

Meanwhile Cornell 2010 has two BCS true road wins (Alabama, St. John's) and then two Top 25 blowout wins (Temple, Wisconsin). Cornell also had wins over Vermont (NCAA participant) and true road wins at A-10s UMass, La Salle as well as the home A-10 win vs. St. Joes.

Not only does Cornell's resume have the edge, the comparison is not even close.