Let's Face It - We are the Kings of November and December

[freshmanjs]

Again, you're being subjective. Personally I'd characterize many of these seasons differently. For example, we were the pre-season #1 team in the country in 1988-89 -- how could reaching the Final Four and losing the semifinal be considered exceeding expectations? And I think a lot of people around here would argue we fell short in 1999, and I doubt many would consider losing to a lower seeded directional school in 1996 should be considered "on target." In 1998 we fell short just as much as we did in 2009; in 2000 we fell short as much as we did in 2011; and in 2005 and 2008, while we lost to lower seeds, I personally didn't expect much (if at all) better than we did. So, in my subjective opinion, even using your arbitrary cutoff, the years from 1995 to 2004 included four or five "fall shorts" (depending on how you view 1999) and four or five "on targets." The years from 2005 to 2012 included four or five "fall shorts" (depending on how you view 2005) and three or four "on targets" or "exceeded." Not so much of a difference.

Also, you're completely ignoring 1981 to 1985 (or at least 1984 to 1985 if all you care about is the NCAAT), but we "fell short" in those years too, and they're "pre 2004," right?

Your earlier post spoke about pre-2001 vs. post and now you're talking about pre-2004 vs. post, so to me it seems like you're moving the goalposts again. I know you've mentioned small sample size, but the later you go with your cutoff date the smaller the sample becomes. At this point you're comparing 24 years to 8, which doesn't make a lot of sense to me.

i'll try to address your points.

1) you can re-do everything i did with a non-arbitrary measure of expected wins per seed. if you do that, my point still holds

2) i disagree on your assessment of many of these seasons.

1988 - 89 - now you are the one moving the goalposts. preseason ranking has nothing to do with what i said at all. duke was a 2 seed and got to the final 4. that is not falling short of expected performance for seed.

1999, we won more than expected number of games for seed

1996, we lost an 8-9 game. those are essentially toss-ups so i said on target.

1998, we won 3 games as a #1 seed - about typical performance.

2009, maybe -- i could see a case for on target there vs fell short.

2000 vs 2011, i think are very different cases because of the dunleavy illness as i pointed out. you could say the return of kyrie offsets that. maybe. but i don't think so. duke was a #1 seed without kyrie and had the option to play without him if that was best for the team.

2005 and 2008 - if the discussion is how did the team do in the ncaat vs kedsy's ingoing expectations, that is a) totally uninteresting to me and b) not something i have any ability to evaluate. both of these teams did not win the expected number of games for seed

3) ok -- duke fell short twice before 1986. noted. does not change my conclusion

4) i see very little difference between 2001 cutoff and 2004 cutoff in terms of the conclusion. if the fact that i mentioned both is a problem for you, let's stick wiht 2004. i think the difference more striking after then.

5) yes!!! i am comparing 24 years to 8. and the 8 have not been at the same level in ncaa performance relative to seed (which reflects regular season). i am interested in a discussion about why that is. is it random, luck, something else? that would be a great conversation to have but for some reason you insist on denying that the 8 years have been different.

[throatybeard]

1996, we lost an 8-9 game. those are essentially toss-ups so i said on target.

Indeed, I think since 1985, the Nines have a winning record over the Eights. Or at least that was the case for many years.

Also, with respect to the Eastern Michigan game, virtually the whole team was hurt in one fashion or another.

[bedeviled]

1) you can re-do everything i did with a non-arbitrary measure of expected wins per seed.

By nature, expectation is subjective, thus not a non-arbitrary measure.

2) i disagree on your assessment of many of these seasons. 1988 - 89 - now you are the one moving the goalposts. preseason ranking has nothing to do with what i said at all. duke was a 2 seed and got to the final 4. that is not falling short of expected performance for seed.

I realize you are discussing this with Kedsy, but, to me, this IS the point. Just like choosing different timeframes, choosing a particular relative measures is subjective. (falling short of expected performance...of a seed, of preseason ranking, coach's poll, season win%, RPI, Pomeroy, of my own personal desires, etc.) I think, upthread, Kedsy compared the actual outcomes for comparative timeframes. If those outcomes didn't meet a particular set of expectations, then perhaps the problem is with the expectations.

5) yes!!! i am comparing 24 years to 8. and the 8 have not been at the same level in ncaa performance relative to seed (which reflects regular season). i am interested in a discussion about why that is. is it random, luck, something else? that would be a great conversation to have but for some reason you insist on denying that the 8 years have been different.

After y'all establish the ground rules, I will be interested in hearing your reasoning for what you think is the difference in those years.

[freshmanjs]

By nature, expectation is subjective, thus not a non-arbitrary measure

i meant compare number of ncaat wins actually achieved to the average number of ncaat wins achieved by previous teams with the same seed in the 64+ team era. there is a table upthread with those numbers.

[bedeviled]

I get that and saw the table upthread. However, seeds are based on subjective expectations built on history, personal preference, ideas of what conferences/schedules are harder, etc.
EDIT: Those factors are not necessarily controlled / balanced across teams (like Duke!)

[freshmanjs]

I get that and saw the table upthread. However, seeds are based on subjective expectations built on history, personal preference, ideas of what conferences/schedules are harder, etc.

i'm with you. it is possible that duke has been over-seeded in this recent period. that is one of several possible explanations i have in mind for the recent 8 year period. there is a counter balancing factor to consider, though, which is that over-seeding buys you easier opponents. #1 seeds do well in the tournament both because they are good teams AND because they have easier games.

[OldPhiKap]

Any calculation based upon the assumption that a #1 seed is "supposed" to go to the final four, and that a #2 seed is "supposed" to get to the regional championship game, has little basis in reality and is fundamentally flawed. With all due respect, that simply ignores the nature of a six-game, one-loss elimination tournament. It also ignores the fact that, in each bracket, each of the top four teams are pretty darn good and in fact each of the teams that got there have a proven record sufficient to merit the invitation or earned bid. The seeding itself is highly subjective to start.

And to my mind, if we played out hearts out and lost -- no matter what round -- that is not below my expectations.

If someone else feels differently, well, we are all entitled to our subjective views. But they are exactly that.

[freshmanjs]

Any calculation based upon the assumption that a #1 seed is "supposed" to go to the final four, and that a #2 seed is "supposed" to get to the regional championship game, has little basis in reality and is fundamentally flawed. With all due respect, that simply ignores the nature of a six-game, one-loss elimination tournament. It also ignores the fact that, in each bracket, each of the top four teams are pretty darn good and in fact each of the teams that got there have a proven record sufficient to merit the invitation or earned bid. The seeding itself is highly subjective to start.

And to my mind, if we played out hearts out and lost -- no matter what round -- that is not below my expectations.

If someone else feels differently, well, we are all entitled to our subjective views. But they are exactly that.

why do we talk on this board about "subjective" views as if there is something wrong with having subjective views. very odd.

personally, i don't have a view of "supposed to." but, i do have a view of typical performance of a team with a given seed and that is, to me, useful as a benchmark for comparison. fine if you disagree - feel free to ignore it.

fwiw, i agree about the play their hearts out point with respect to my expectations of the people on the team.

[Wander]

Any calculation based upon the assumption that a #1 seed is "supposed" to go to the final four, and that a #2 seed is "supposed" to get to the regional championship game, has little basis in reality and is fundamentally flawed.

No one here has been using that assumption.

[bedeviled]

why do we talk on this board about "subjective" views as if there is something wrong with having subjective views.

It's likely related to the passion involved for AND against Duke basketball. It's a good thing. Truthfully, nearly all posts are opinions or subjective, but it is nice to try to distinguish "fact" from "feeling" when we can, especially when there is the potential of starting or perpetuating false negative Duke basketball memes. Take a look back at your post #69 as an example of a post that claims "fact" but could have been phrased a number of different, more palatable ways. The thread has many thoughtful, reasoned arguments for and against the claim, including the actual results pointed out by Kedsy and tourney performance of Blue Bloods by Devil84.
Most of all, I, personally, find nothing more annoying in a forum than the "Pudding Pops > Matt Ryan. FACT!" comments.

[Duvall]

Indeed, I think since 1985, the Nines have a winning record over the Eights. Or at least that was the case for many years.

Also, with respect to the Eastern Michigan game, virtually the whole team was hurt in one fashion or another.

And the only NBA player on the floor wasn't wearing a Duke uniform.

Actually, looking back at some of these upsets with a better sense of the talent on both sides, it seems like the real scandal in these games is that the opposing coaches couldn't get more wins out of their elite talent during the regular season. Florida in 2000 and LSU in 2006 were particularly egregious.

[freshmanjs]

It's likely related to the passion involved for AND against Duke basketball. It's a good thing. Truthfully, nearly all posts are opinions or subjective, but it is nice to try to distinguish "fact" from "feeling" when we can, especially when there is the potential of starting or perpetuating false negative Duke basketball memes. Take a look back at your post #69 as an example of a post that claims "fact" but could have been phrased a number of different, more palatable ways. The thread has many thoughtful, reasoned arguments for and against the claim, including the actual results pointed out by Kedsy and tourney performance of Blue Bloods by Devil84.
Most of all, I, personally, find nothing more annoying in a forum than the "Pudding Pops > Matt Ryan. FACT!" comments.

i do have a pretty strong view that the facts back up a difference in ncaat performance for duke in the recent 8 years vs the period that came before. there is obviously interpretation applied to those facts to reach that conclusion. Kedsy himself acknowledges (not in a manner that makes it at all enjoyable to be part of this board) over and over again that the 1986-1994 period was superior performance. the blue blood performance analysis is very interesting but does not comment on the comparison of duke pre 2001 (or 2004) to after.

[bedeviled]

I don't share your view as stated - I think our recent performance has been on par with our near-recent performance. But, I do believe that our performance has been worse compared to expectations based on tournament seeding recently, which appears to be your supported assertion. Thus, if you believe that this is a systematic trend, I remain curious about your evidence of a systematic cause on the other side of the equation. Furthermore, I am interested in whether or not that systematic cause is meaningful (isolated to Duke, can be adjusted, etc).
The primary reason for this post, though, is to clarify my last post. I should have included the disclaimer that I do LOVE Pudding Pops, I just have a difficult time comparing them to Matt Ryan.

[OldPhiKap]

No one here has been using that assumption.

I guess I am unclear what "expected number of wins relative to seeding" would mean other than that. But it may be my misunderstanding.

How many games is a #1 seed "expected" to win?

[Kedsy]

here is how i'd assess ncaat relative to regular season

yes!!! i am comparing 24 years to 8. and the 8 have not been at the same level in ncaa performance relative to seed (which reflects regular season).

Before, you were talking about "ncaat relative to regular season," now you want to talk about underperforming seed. That's fine, but as someone said earlier, there are many ways to "reflect" the regular season. I would argue seeding in the NCAAT is not a particularly good way to do it, however. Be that as it may:

1984 - ??? (3 vs. 6 2nd round loss) -- should be fell short
1985 - ??? (3 vs. 11 2nd round loss) -- should be fell short
1986 - on target (1 vs. 7 made FF)
1987 - on target (5 vs. 1 S16 loss)
1988 - exceeded (2 vs. 1 made FF)
1989 - exceeded (2 vs. 1 made FF)
1990 - exceeded (3 vs. 1 made FF)
1991 - exceeded (2 vs. 1 made FF)
1992 - on target (1 vs. 2 made FF)
1993 - fell short (3 vs. 6 2nd round loss)
1994 - exceeded (2 vs. 1 made FF)
1995 - n/a
1996 - on target (8 vs. 9 1st round loss) -- should be fell short
1997 - fell short (2 vs. 10 2nd round loss)
1998 - on target (1 vs. 2 E8 loss) -- should be fell short
1999 - on target (1 vs. 6 made FF)
2000 - on target (1 vs. 5 S16 loss) -- should be fell short
2001 - on target (1 vs. 6 made FF)
2002 - fell short (1 vs. 5 S16 loss)
2003 - on target (3 vs. 2 S16 loss)
2004 - on target (1 vs. 7 made FF)

2005 - fell short (1 vs. 5 S16 loss)
2006 - fell short (1 vs. 4 S16 loss)
2007 - fell short (6 vs. 11 1st round loss)
2008 - fell short (2 vs. 7 2nd round loss)
2009 - fell short (2 vs. 3 S16 loss)
2010 - exceeded (1 vs. 3 made FF)
2011 - fell short (1 vs. 5 S16 loss)
2012 - fell short (2 vs. 15 1st round loss)

In parenthesis I have included either our last game or the game that got us into the Final Four.

First of all, if you're strictly talking about comparing performance against seed, you can't count 2 beating 1 as "exceeding" and 1 losing to 2 as "on target." You can't say 8 losing to 9 is "on target." You can't make injury excuses for one #1 losing to a #5 and not another. If you change 1996, 1998, and 2000 to "fell short," and it really does look like cherry picking.

i am interested in a discussion about why that is. is it random, luck, something else? that would be a great conversation to have but for some reason you insist on denying that the 8 years have been different.

Actually, what I've been insisting is that the years from 1986 to 1994 were different. The years before 1986 and after 1994, not so much.

We only have five post-1994 "on target" years. Here is our NCAAT path to the Final Four (or to a loss before the Final Four) for those five years:

1999: 16-9-12-6
2001: 16-9-4-6
2003: 14-11-2 (loss)
2004: 16-8-5-7
2010: 16-8-4-3

At the outset, why are 2003 and 2004 in the "we were good" days and 2005 and 2006 in the "something happened to us" days? That doesn't sound arbitrary to you?

Putting that aside, during all those years prior to the Final Four we only beat one team seeded higher than 4th, and that was in 2010 (during the "down" time). Other than 2010, prior to the Final Four we only beat two teams total seeded higher than 6th.

If the 2003 team had played a second round opponent as tough as 2009 West Virginia (instead of 11th seeded Central Michigan), would they have performed any better than our 2008 team? If 2006 LSU hadn't won on a last-second prayer to beat #12 Texas A&M, would our 2006 team fared just as well as the 2004 team? You may not be willing to entertain those sorts of questions as part of this debate, but when you have so few data points that luck factor makes the whole thing. Because if 2006 and 2008 are even with 2003 and 2004, then you're left with two Final Four visits by Shane Battier against the 2010 team, and to say that's a trend is silly.

Again, I say it ultimately comes down to what was different about 1986 to 1994, not what's been different about 2005 to 2012.

[davekay1971]

I guess I am unclear what "expected number of wins relative to seeding" would mean other than that. But it may be my misunderstanding.

How many games is a #1 seed "expected" to win?

I think they were using a more reasonable calculation of average number of tournaments wins for each seed. i.e.: over a given period of time, 1 seeds win, on average, x.yz number of games per tournament. Then they compared Duke's performance to that and, if Duke won less games at a given seen than average, it was termed underperforming. Given the sample size (i.e.: Duke's number of times over a given period as a 1, 2, 3, etc seed), I'm having a hard time reaching any significant conclusions. I simply don't have the inclination to do a chi-square or linear regression or z-statistic or whatever would apply here, but I feel safe in assuming that, with such a small sample size, our standard deviation is pretty huge and Duke's performance over the last 10 years would be the perfect definition of a statistically insignificant trend.

However, accepting the argument that Duke has underperfomed the expected number of wins for our given seed over the last 10 years, I would offer a few possible explanations

1) It's a small sample size and we're simply seeing random variation (i.e.: kids have bad games, kids on other teams have great games, injuries, dumb luck, etc)
2) Coach K forgot how to coach the NCAA tournament
3) Duke may have a tendency to be overseeded given the program's, and K's, well earned tournament reputation
4) The greater dispersal of talent and quality of coaching has made the tournament a more fundamentally random event.
5) The NCAA's seeding system sucks and underrates less well established programs like Lehigh, VCU, etc. Frankly, both Lehigh and VCU were criminally underrated when they bounced us.

I personally think reason number 2 is complete bunk and the other 4 probably are all factors...particularly number 1.

[freshmanjs]

I don't share your view as stated - I think our recent performance has been on par with our near-recent performance. But, I do believe that our performance has been worse compared to expectations based on tournament seeding recently, which appears to be your supported assertion. Thus, if you believe that this is a systematic trend, I remain curious about your evidence of a systematic cause on the other side of the equation. Furthermore, I am interested in whether or not that systematic cause is meaningful (isolated to Duke, can be adjusted, etc).
The primary reason for this post, though, is to clarify my last post. I should have included the disclaimer that I do LOVE Pudding Pops, I just have a difficult time comparing them to Matt Ryan.

i have a lot of hypotheses and no way to prove them. i also think there is a decent chance that this is just random variation from small sample size, as i've stated many times in this thread.

some ideas that i think are worth exploring (although some here seem to want to cut me off before this point and deny permission to discuss this topic):

1) the roll of constant favorite takes a toll. for some reason playing as a #1 seed or even the overall #1 seed has made Duke more tentative in the NCAA-T vs prior periods when Duke was actually seeded lower but had better ncaa-t performance. some of the great runs included somewhat surprising wins and I think coach K was at his very best preparing for games we were expected to lose (Purdue, Temple, Georgetown, UNLV 2). we haven't had many games we were expected to lose in the ncaat recently.

2) perhaps the change in quality of other ACC teams for the worse has made Duke less prepared for the NCAA-t than back in the 88-94 timeframe when the conference was so good. duke's sos is usually quite good as evaluated by the rpi. however a lot of this is because we don't play cupcakes and not because we play a lot of really tough teams. this varies by year and i'd love to see a history of our record vs top 10, 25, 50 teams by year going back to 1986. i don't have the data. i think this has also coincided with fewer really tough home/home non conference games (michigan, arizona, ucla, lsu (shaq) vs more recently st johns).

3) coach k has actually gotten better overall as a coach. this has led, in particular, to the incredibly good early season records. couple that with his recent scheduling philosophy which has seemed to be designed to optimize RPI, and maybe there is an over-seeding phenomenon happening with Duke

4) personnel. recent duke teams have not had the ingredients that the great coach k teams had. now, i'm not talking about absolute talent level here. obviously, overall talent level for duke and for everyone else was much higher 20 years ago than it is today. what i mean, though is on a relative basis, we have not had key ingredients. we have not had balance between perimeter and post scoring in many of the recent seasons and we have not had a relatively great point guard in many of these seasons. contrast to the earlier period when we had both of those items. i think when it comes to ncaa-t play, it is very difficult to win without both.

5) a lot of this is really a reflection of one recruiting class that just didn't work out. paulus, mcroberts, pocius, boateng, boykin. i like them all, especially paulus. but this class was not the juggernaut that it was predicted to be. this led directly to several years of the issues in #4.

6) coach k has changed his priorities, putting more emphasis on each regular season game and the ACCT than he used to. it is REALLY interesting that our ACCt performance was so good during this timeframe. maybe that wears down the team as well.

anyway, i'm not sure i believe any of these. also to state it again because i expect some folks to bite my head off about it -- i am not at all sure there is a trend here beyond randomness.

[freshmanjs]

At the outset, why are 2003 and 2004 in the "we were good" days and 2005 and 2006 in the "something happened to us" days? That doesn't sound arbitrary to you?

Again, I say it ultimately comes down to what was different about 1986 to 1994, not what's been different about 2005 to 2012.

no, including our 2004 final 4 team in a list of good years does not seem arbitrary to me.

we must be talking past each other totally. "we were good" in ALL of these eras except for 1995-96. and maybe i'd say the 07 team was not very good. i would not remotely agree, AT ALL, with characterizing these eras in to "we were good" and "something happened to us." not at all. not remotely close to what i think.

what i'm talking about is different NCAA-T performance.

[freshmanjs]

I think they were using a more reasonable calculation of average number of tournaments wins for each seed. i.e.: over a given period of time, 1 seeds win, on average, x.yz number of games per tournament. Then they compared Duke's performance to that and, if Duke won less games at a given seen than average, it was termed underperforming. Given the sample size (i.e.: Duke's number of times over a given period as a 1, 2, 3, etc seed), I'm having a hard time reaching any significant conclusions. I simply don't have the inclination to do a chi-square or linear regression or z-statistic or whatever would apply here, but I feel safe in assuming that, with such a small sample size, our standard deviation is pretty huge and Duke's performance over the last 10 years would be the perfect definition of a statistically insignificant trend.

However, accepting the argument that Duke has underperfomed the expected number of wins for our given seed over the last 10 years, I would offer a few possible explanations

1) It's a small sample size and we're simply seeing random variation (i.e.: kids have bad games, kids on other teams have great games, injuries, dumb luck, etc)
2) Coach K forgot how to coach the NCAA tournament
3) Duke may have a tendency to be overseeded given the program's, and K's, well earned tournament reputation
4) The greater dispersal of talent and quality of coaching has made the tournament a more fundamentally random event.
5) The NCAA's seeding system sucks and underrates less well established programs like Lehigh, VCU, etc. Frankly, both Lehigh and VCU were criminally underrated when they bounced us.

I personally think reason number 2 is complete bunk and the other 4 probably are all factors...particularly number 1.

has anyone ever suggested that #2 is a reasonable possible explanation?

[Kedsy]

no, including our 2004 final 4 team in a list of good years does not seem arbitrary to me.

OK, I'll agree we're talking past each other. Absent something major happening at that juncture (e.g., a change in conference or a coaching change), saying that something that happened in 2004 is part of a 24 year trend but something that happened in 2005 is not but instead is part of a more recent 8 year trend seems to me to be the height of arbitrariness.