SI's Stewart Mandel - fair comment

[FerryFor50]

How many of those folks do you think were in Rupp Arena when unc played there? Or, even better, how many of Duke's OOC opponents wouldn't feel their chances for a win had increased greatly if they played the blue devils in charlotte, or greensboro, or even in the meadowlands, rather than have to play them at Cameron? There is a HUGE difference between playing near somebody's home floor and playing in their regular home arena -- especially when it's on-campus. Let's be honest about this.

There is.

But Duke plays a lot of away games in Maui, New Jersey/New York, etc. Just because a lot of Duke fans live in the north doesn't mean Duke shouldn't be allowed to play there. And last time they did, who did they play? Pitt.

They don't play all their road games in Greensboro.

UNC played true road games, but against lesser teams.

[wiscodevil]

Nice - real nice. "Yapping". I take it you are the standard bearer of civility around here huh? You certainly aren't the leader in reading comprehension or forensics.

you're kidding right?

[devildownunder]

Nice - real nice. "Yapping". I take it you are the standard bearer of civility around here huh? You certainly aren't the leader in reading comprehension or forensics.

The title of the article is:

Dissecting Duke: Why the Devils have sunk down stretch since 2001

This is a bit away from what you were getting at but...


Most people who are not in media probably don't know that reporters write probably less than one tenth of 1% of the headlines you see on their stories, no matter what medium they are in, unless you're reading a blog, where one person writes every single thing you see. In this case, I think the headline, while provocative, is a little misleading, but because he almost certainly didn't write it, I'm not going to blame him for it. Mandel touches on duke's second-half record vs. it's first-half one but he seems to address that in the context of discussing what he considers to be a weakening of the earlier-season schedules. The thrust of the article seems to be that duke has had gaudy records and seedings in the last handful of years but has failed to live up to them. In essence, he is calling us something of a paper tiger. I think a more accurate headline would have said something along those lines.

[devildownunder]

There is.

But Duke plays a lot of away games in Maui, New Jersey/New York, etc. Just because a lot of Duke fans live in the north doesn't mean Duke shouldn't be allowed to play there. And last time they did, who did they play? Pitt.

They don't play all their road games in Greensboro.

UNC played true road games, but against lesser teams.

I have no problem with duke playing in New Jersey/NY once or twice a year. No problem with Maui/alaska either. In fact, I applaud participation in those tournaments because they usually lead to games against quality competition.

But they are not true road games, anymore than playing in jersey is a true home game.

Also, nothing about those contests precludes putting two major, anticipated top 20 home-and-homes on the schedule. Real home-and-homes, not just in the same state or whatever. And on this point, I agree with mandel 100% and I'm glad he put the numbers there to show it. Duke simply does not try to schedule these games anymore and I know I can't prove it but I think it is hurting us. I don't recall unc's entire schedule and it doesn't matter. This isn't about what we've done relative to them, it's just about us.

[FerryFor50]

I have no problem with duke playing in New Jersey/NY once or twice a year. No problem with Maui/alaska either. In fact, I applaud participation in those tournaments because they usually lead to games against quality competition.

But they are not true road games, anymore than playing in jersey is a true home game.

Also, nothing about those contests precludes putting two major, anticipated top 20 home-and-homes on the schedule. Real home-and-homes, not just in the same state or whatever. And on this point, I agree with mandel 100% and I'm glad he put the numbers there to show it. Duke simply does not try to schedule these games anymore and I know I can't prove it but I think it is hurting us. I don't recall unc's entire schedule and it doesn't matter. This isn't about what we've done relative to them, it's just about us.

Well, the point Coach K has made is... are NCAA tourny games really TRUE road games? No.

So there are just 2 schools of thoughts on the matter... but it's not why Duke "fades" at the end of a season.

[devildownunder]

Good point. It's kind of tough to have upside when you're a #1 seed. Especially when you're a #1 seed overall.

Correct, but Mandel isn't berating duke for lack of upside. He is saying that Duke has been getting those No.1 and No. 2 seeds consistently for years, only in the last 5 or 6 years, we haven't been living up to them, whereas before, more often than not, when we had those seeds, we were good enough to live up to them. His argument is concerned with determining why that may be the case.

[FerryFor50]

Correct, but Mandel's isn't berating duke for lack of upside. He is saying that Duke has been getting those No.1 and No. 2 seeds consistently for years, only in the last 5 or 6 years, we haven't been living up to them, whereas before, more often than not, when we had those seeds, we were good enough to live up to them. His argument is concerned with determining why that may be the case.

What he's failing to mention is that, though Duke gets a #1 or #2 seed, they constantly get put in the toughest brackets of the tournament. #1 seed isn't doing you much good when your road is the toughest.

[devildownunder]

Well, the point Coach K has made is... are NCAA tourny games really TRUE road games? No.

So there are just 2 schools of thoughts on the matter... but it's not why Duke "fades" at the end of a season.

No, they aren't true road games, although they are close to such for duke far more often than for most teams.

Games at home against midmajors who have lost four senior top-line players from their final four team of the previous season -- without a rematch in their gym -- aren't similar to NCAA games either, but we scheduled George Mason a couple of years ago. So there are more reasons to schedule games than just whether or not they precisely simulate NCAA tourney conditions. Taking a team into an opponent's home gym toughens and galvanizes them.

Look, nothing anyone says here is ever going to change K's mind on anything but the fact is that our tournament success on the whole, as far as how well we did at living up to our seeding, was much better when K took the boys into the lion's den a few times. That doesn't mean one led to the other, but it doesn't support the argument that we don't need these games either.

[devildownunder]

What he's failing to mention is that, though Duke gets a #1 or #2 seed, they constantly get put in the toughest brackets of the tournament. #1 seed isn't doing you much good when your road is the toughest.

How exactly do you support that statement? On a couple of occasions, duke has been the no.1 overall seed, meaning the committee would have put us in what it thought was the weakest bracket, with the weakest 2, strongest 3, weakest 4, etc. meaning all the teams in our half would be the weakest in the field at their seeding.

I am at a loss to think of what evidence you could have that duke was consistently placed in the toughest regional.

[FerryFor50]

How exactly do you support that statement? On a couple of occasions, duke has been the no.1 overall seed, meaning the committee would have put us in what it thought was the weakest bracket, with the weakest 2, strongest 3, weakest 4, etc. meaning all the teams in our half would be the weakest in the field at their seeding.

I am at a loss to think of what evidence you could have that duke was consistently placed in the toughest regional.

Well, I have no evidence, of course, just opinion.

I just look at the brackets from years to year and say to myself "wow, what did we do to deserve THAT?"

If I find some time, I'll see if I can dig up old brackets and look at the matchups in hindsight... but I just know that I notice that trend every year.

For instance, when Florida won their first NC... they had one of the easiest roads to the finals I can remember (both by "luck" of the draw and other teams losing before they had to face Florida).

The NCAA tourny is really a flawed tourny because the best team doesn't always win. It's mostly luck of the draw. It's great entertainment, though, because you get to see upsets and "one shining moment" but only when you have 5 game series do you find out the best teams. (Not saying college should adopt that, because it makes no sense.)

[devildownunder]

Well, I have no evidence, of course, just opinion.

I just look at the brackets from years to year and say to myself "wow, what did we do to deserve THAT?"

If I find some time, I'll see if I can dig up old brackets and look at the matchups in hindsight... but I just know that I notice that trend every year.

For instance, when Florida won their first NC... they had one of the easiest roads to the finals I can remember (both by "luck" of the draw and other teams losing before they had to face Florida).

The NCAA tourny is really a flawed tourny because the best team doesn't always win. It's mostly luck of the draw. It's great entertainment, though, because you get to see upsets and "one shining moment" but only when you have 5 game series do you find out the best teams. (Not saying college should adopt that, because it makes no sense.)

It's pretty easy to find fans of any team that is consistently in the field of 64 (cbs and the ncaa bedeviled! it's 64!) who consistently think their team has received a right royal raw deal from the selection committee. So while I certainly can't refute your claims, I can say that duke fans don't have any monopoly on the sentiment.

And I disagree that the tournament is ALL about luck. Winning it all in any particular year generally does require some luck, such as a great escape, or 1 or 2 favourable matchups, but I think making a consistently strong, or weak, showing has a lot to do with coaching and preparation and knowing how to get your team ready under the circumstances. The relevant decisions involve several variables in play throughout the season and possibly even the offseason, not just right at tournament time. And I think one of those variables is scheduling.

[BD80]

[QUOTE=devildownunder;107929]

The title of the article is:

Dissecting Duke: Why the Devils have sunk down stretch since 2001

This is a bit away from what you were getting at but...


I think the headline, while provocative, is a little misleading, but because he almost certainly didn't write it,

So the editor didn't understand the article?

What about these paragraphs?

If you've followed the Blue Devils closely over the past seven years, however, this week's sudden swoon should come as little surprise. If anything, the scenes described above probably seemed extremely familiar, if not inevitable, considering where they fell on the calendar.

A closer look at the numbers, however, reveals Duke has not been nearly as dominant late in the season. The Devils' demise has not been limited solely to the Big Dance, either -- in most cases, their swoon begins in February.

I don't think we differ that much in our interpretation. Mandel discusses unmet expectations in that prior to 2002, we had continued our early season successes. I find the focus of the article is the reasons we are "swooning" beginning each February - implying that our current two game losing streak is a swoon.

[sagegrouse]

Maybe this demonstrates a lack of understanding on my part but aren't the supporting facts in anyone's argument selective? Mandel starts looks at a trend that has developed over the last 6 or 7 years, coming right up to the present day. He identifies a trend, then offers some ideas on some of the factors that may be behind that trend. Anyone else is certainly free to try to come up with their own stats but how else can you possibly make any argument? And if he doesn't use any stats he will be criticised for that, correct?

It seems to me that in analyzing the argument it is unfair and unproductive to slam someone for attempting to support their statements or for being "selective". Our job is to weigh the quality of the argument, based on the validity of the premise and quality of the supporting details, both in light of what is there and what isn't. In this case, I think the writer's premise and details are drawn from facts and his (rather cautious, i think) conclusions are reasonable. That doesn't mean he has come up with a universal truth but I do think that dismissing his case as irrelevant because it is "flawed" is closed-minded.

Part of the problem is that he falsified the data AND the analysis (failing to eliminate current college athletes when calculating the % making it in the NBA is worse than lying).

It appears to be the case that the earlier years were unusually productive for Duke McD AA's -- not that the recent ones have done bad.

And, no, when doing analytical research, you don't start with your conclusions (Duke's FF success the past six years is not as good as the previous 15) and try to force-fit and explanation. As others have noted, Duke's hasn't done badly compared to other teams -- only really in comparison to PAST Duke teams. And Duke McD AA's since 2003 have done about as well as the average for all McD AA's, especially if you consider that the sample is dominated by first and second year players.

sagegrouse

[devildownunder]

[QUOTE=BD80;107966]

So the editor didn't understand the article?

What about these paragraphs?

If you've followed the Blue Devils closely over the past seven years, however, this week's sudden swoon should come as little surprise. If anything, the scenes described above probably seemed extremely familiar, if not inevitable, considering where they fell on the calendar.

A closer look at the numbers, however, reveals Duke has not been nearly as dominant late in the season. The Devils' demise has not been limited solely to the Big Dance, either -- in most cases, their swoon begins in February.

I don't think we differ that much in our interpretation. Mandel discusses unmet expectations in that prior to 2002, we had continued our early season successes. I find the focus of the article is the reasons we are "swooning" beginning each February - implying that our current two game losing streak is a swoon.

Mandel does come at it from that angle at the start. But, in the main, the article stacks up as a case for our recent teams being paper tigers, saying that they were never that good to begin with, rather than that they "swooned", which to me implies a team playing beneath its abilities. That may be just a more expansive way of expressing the same idea.

[ArtVandelay]

This is an utterly ridiculous piece wrapped in the integument of falsified statistics.

First: What's his point? Is is that Duke has been totally overrated since 2002 because the players are no good? Or that the weak non-conference schedule doesn't prepare the team for the NCAAs? It's not clear either to me or to the author. And, of course, for three decades columnists were opining that the tough ACC tournament made it impossible to win the NCAA championship. The Duke-Duke-UNC run from 1991 to 1993 muted those voices.

Second: He fudged the data. Of the 17 Duke McD's, nine are still in college (53%). Of the 159 from other schools, only 60 are still in college (38%)! Looking at the proportion in the NBA of those having left college, Duke has four of eight (according to his data). The others are 60 out of 99 (61%). This is not a statistically meaningful difference, if his data were correct -- which they aren't. There are actually five Duke NBA roster players from the McD sample since 2003 -- Luol, JJ, Shelden, Shav and.... Josh, who is listed on the Trailblazers roster.

WRT to the number of starters lets note that Shav is injured and the others (except Luol) are in their first two years.

Third: Probably the biggest problem from a statistics standpoint (he is not a professional statistician, and I wouldn't recommend it) is that he picked the pinnacle of 2001 and 2002 and then used it to compare with the (inevitable) valley that followed. Battier, Boozer, Dunleavy, JWill (if uninjured) were tremendous NBA players and Duke went 4 for 4 in turning McD AA's into NBA starters. This is selectivity bias of the worst kind. For example if used some fixed time interval (2000s or last ten years), he would get a totally different answer.

The reasonable conclusion from the data I see is that the McD AA's who left Duke between 1999 and 2002 did phenomenally better than their peers who went to other schools. The ones since than have done just about average -- that is, average for McDonald All-Americans.

Ugh! Double ugh!

sagegrouse

These critiques pretty much nail it, in my view, although I wouldn't say the end result is "utterly ridiculuous." His column is relatively "fair," in that he doesn't make stuff up out of whole cloth. It's just misleading, because he has presented the statistics in a very specific way to make a very dubious point. One of the very first comments on this thread was that you can say this for just about every team, except for the part about the McDonald's AAs. That doesn't make his points untrue, but it does make you wonder why Duke was chosen as the subject, as opposed to some other team.

And as for the part about the Mickey Ds, as Sagegrouse points out, he cherry-picked the period from 2002-2007, during which a number of our AAs are still in school, while others have suffered pretty major injuries. Really it's just the 2004 class (Paulus, McBob, et al) that turned out to be a pretty major bust and seems to continue to cast a shadow on the program for some unknown reason. Again, it's not that this criticism that Duke AAs are not NBA starters is WRONG, but it's a) misleading, and b) has some very reasonable explanations. The fact that it's both of those things also makes you wonder about the motivations of the author.

[devildownunder]

Part of the problem is that he falsified the data AND the analysis (failing to eliminate current college athletes when calculating the % making it in the NBA is worse than lying).

It appears to be the case that the earlier years were unusually productive for Duke McD AA's -- not that the recent ones have done bad.

And, no, when doing analytical research, you don't start with your conclusions (Duke's FF success the past six years is not as good as the previous 15) and try to force-fit and explanation. As others have noted, Duke's hasn't done badly compared to other teams -- only really in comparison to PAST Duke teams. And Duke McD AA's since 2003 have done about as well as the average for all McD AA's, especially if you consider that the sample is dominated by first and second year players.

sagegrouse

Mandel says that the guys we had before were the real deal and the guys we've had since were not. He uses this as evidence that the public's opinion of duke's talent/ability level over the last few years has been higher than the actual reality of these duke teams.

From Mandel:

A widely held assumption among college basketball followers is that Duke, being Duke, basically gets its choice of the nation's top recruits every year. How do we know this? Because Duke always signs the most McDonald's All-Americans.

Currently, there are eight McDonald's alums on the Blue Devils' roster. That's right -- eight. The next-closest team, North Carolina, has five.

McDonald's honorees come in all shapes and sizes, however -- many become NBA All-Stars (Carmelo Anthony, Chris Paul, LeBron James) or solid starters (Raymond Felton, Mo Williams, Daniel Gibson). Others (DeAngelo Collins?) fall off the face of the earth. Most, however, fall somewhere in between: Solid college players, some of whom do (Glen Davis, David Harrison) or don't (Chris Thomas, Travis Garrison) make it to the next level.

In the post-Elton Brand/Shane Battier/Carlos Boozer era, most of Duke's McDonald's All-Americans have fallen into the latter two categories. ...



Whatever he did with those statistics you cited, his point that the mcdaa's in recent years haven't been the studs that the mcdaas from other years were is still valid, unless you really want to argue that guys like boateng and mcroberts are on a par with boozer and brand.

So what is your point about the story? Not trying to be difficult, just want to understand? Are you saying the rest of his story is invalid because he didn't handle these stats correctly? What's your assertion?

[devildownunder]

These critiques pretty much nail it, in my view, although I wouldn't say the end result is "utterly ridiculuous." His column is relatively "fair," in that he doesn't make stuff up out of whole cloth. It's just misleading, because he has presented the statistics in a very specific way to make a very dubious point. One of the very first comments on this thread was that you can say this for just about every team, except for the part about the McDonald's AAs. That doesn't make his points untrue, but it does make you wonder why Duke was chosen as the subject, as opposed to some other team.

And as for the part about the Mickey Ds, as Sagegrouse points out, he cherry-picked the period from 2002-2007, during which a number of our AAs are still in school, while others have suffered pretty major injuries. Really it's just the 2004 class (Paulus, McBob, et al) that turned out to be a pretty major bust and seems to continue to cast a shadow on the program for some unknown reason. Again, it's not that this criticism that Duke AAs are not NBA starters is WRONG, but it's a) misleading, and b) has some very reasonable explanations. The fact that it's both of those things also makes you wonder about the motivations of the author.

Except that he doesn't make any hateful statements about duke. he doesn't say K is a fraud, he doesn't say the program will never win another NC or anything like that. He basically just says "hey, you probably are consistently hit over the head with 'duke dominates' that you are surprised whenever they lose, but the truth is, they haven't been as dominant as you think they have been, especially late in the seasons, and here are some explanations for why I think this pattern has held up lately."

There is no assault on the program here, that I can see. In a way, I think it actually defends the program. He's saying the team aren't falling apart, they're just not as good as some of their midseason records of late suggest they are. I'd rather be overrated than legitimately underachieve. Just me.

[chrisM]

so to be clear, as a #1 or 2 seed in the tourney, if they don't make the FF or at least elite 8 they have underachieved, right?

We do know what the average team seeded #1 and #2 does (since the NCAA went to the 64 team tourney in 1985 with everyone seeded by a selection committee). I've taken the standard list and removed the wins provided by UNC and Duke during the period I'm studying.

A #1 seed wins, on average, 3.31 games
A #2 seed wins, on average, 2.40 games
A #3 seed wins, on average, 1.78 games
A #4 seed wins, on average, 1.55 games
A #5 seed wins, on average, 1.15 games
A #6 seed wins, on average, 1.30 games
A #7 seed wins, on average, 0.85 games
A #8 seed wins, on average, 0.69 games
A #9 seed wins, on average, 0.58 games
A #10 seed wins, on average, 0.66 games
A #11 seed wins, on average, 0.50 games
A #12 seed wins, on average, 0.50 games
A #13 seed wins, on average, 0.25 games
A #14 seed wins, on average, 0.19 games
A #15 seed wins, on average, 0.05 games
A #16 seed wins, on average, 0.00 games

Obviously, a team can't actually win 3.31 games, but over time those fractional wins should average out.

Let's compare UNC and Duke over the last four years:
2004: Duke seeded 1, expected win total: 3.31 games, actual win total: 4 games
UNC seeded 6, expected win total: 1.30 games, actual win total: 1 game
2005: Duke seeded 1, expected win total: 3.31 games, actual win total: 2 games
UNC, seeded 1, expected win total: 3.31 games, actual win total: 6 games
2006: Duke seeded 1, expected win total 3.31 games, actual win total: 2 games
UNC seeded #3, expected win total: 1.78 games, actual win total: 1 game
2007: Duke seeded #6, expected win total: 1.30 games, actual win total: 0 games
UNC seeded #1, expected win total 3.31 games, actual win total: 3 games

Duke
2004:0.69 games
2005:-1.31 games
2006:-1.31 games
2007:-1.30 games
Total: 3.23 games below average over four years

UNC
2004: -0.30 games
2005: 2.69 games
2006: -0.78 games
2007: -0.31 games
Total: 1.30 games above average over four years

Note that Carolina only out performed their seed average once, but because they came very close every other year, that one year still brings them to above average. Duke also out performed their seed average only once, but even had they won the NC in 2004 they would still be negative (by over a game! 1.23 games, to be exact) because they underperformed so terribly in the other years.

Relative to the average team seeded the same as Duke has been, Duke has done terribly over the past four years. Suggesting that Duke was overrated at the end of the year. Perhaps because of an inflated W/L record thanks to playing a less challenging OOC slate? Was it not playing true OOC road games? Or the players just got worn out because K was using an iron man rotation? Or was Duke just unlucky the past three years in the NCAA tournament?

Chris

[ArtVandelay]

Whatever he did with those statistics you cited, his point that the mcdaa's in recent years haven't been the studs that the mcdaas from other years were is still valid, unless you really want to argue that guys like boateng and mcroberts are on a par with boozer and brand.

So what is your point about the story? Not trying to be difficult, just want to understand? Are you saying the rest of his story is invalid because he didn't handle these stats correctly? What's your assertion?

Again, you're letting the anecdotal evidence overwhelm the objective evidence. You say "Boateng and McRoberts weren't as good as Boozer or Brand." Well sure, but what does that prove? You're overlooking all the other evidence out there and picking one fact that seems to prove your point. For instance, this overlooks that Chris Burgess (in Brand's class) was a tremendous bust as well. Casey Sanders was a pretty big bust in Boozer's class. Granted, the 2004 class was probably a bust all-around, but that doesn't mean that we're not still getting good classes. Perhaps a more fair comparison would be "Is Kyle Singler as good as Boozer?" If Singler stays in school as long as 'Los (3 years), well, maybe. Same with Gerald Henderson or Nolan Smith, maybe. Your argument is essentially, "So what if the facts don't support it?! My point is still valid!"

[ArtVandelay]

Except that he doesn't make any hateful statements about duke. he doesn't say K is a fraud, he doesn't say the program will never win another NC or anything like that. He basically just says "hey, you probably are consistently hit over the head with 'duke dominates' that you are surprised whenever they lose, but the truth is, they haven't been as dominant as you think they have been, especially late in the seasons, and here are some explanations for why I think this pattern has held up lately."

There is no assault on the program here, that I can see. In a way, I think it actually defends the program. He's saying the team aren't falling apart, they're just not as good as some of their midseason records of late suggest they are. I'd rather be overrated than legitimately underachieve. Just me.

Right, which is what makes this, in my view, a more under-handed attack. He isn't explicitly hostile, but it's definitely a hit on the program in my view. I honestly think his underlying point was "Duke is overrated." And I don't think that this underlying point that is particularly accurate, considering that we have a final four appearance in 2004 and were one bad shooting night away from possibly another one in 2006.