As this year’s schedule is finalized and publicized, I wanted to revisit an issue which has been discussed on these boards a couple of times over the years, indicating continued interest here in the topic. The assertion has been made that at some point in his long career at Duke, Coach K had a change in scheduling philosophy, and began to schedule fewer on-campus road games and more neutral-site games in tournament-style arenas, in an effort to prepare the team for the atmosphere it would face in the NCAA Tournament.
I wanted to take a look at our scheduling from a historical perspective and examine a couple of questions. First, has there indeed been a shift over the years in the relative numbers of nonconference on-campus road games, nonconference home games, and neutral-site games at tournament-like arenas? And if so, has the shift in scheduling preferences had any discernible impact on the team’s NCAA tournament performance, as such an impact is assumed to be the goal of the change in scheduling, if such change has occurred in the first place.
OK, methodology. I looked at the schedules from each year beginning in 1983-84, as that was the first team that K took to the NCAA tournament. I counted of course only nonconference games, but I counted only games against what I deemed to be significant or high quality opponents, or even traditional power opponents who happen to be in a down cycle – like teams such as Michigan or St. John’s have been during certain stretches. In other words, I counted games over the years against teams like Washington, Oklahoma, Notre Dame, Temple, and Arizona, for example, but not Stetson, Davidson, Northwestern, ECU, or Harvard. On some of the mid-majors I had to make a judgment call on whether to count them. For each year I tallied the number of these higher quality nonconference opponents we played at home, the number on the road at on-campus arenas, and the number at neutral sites with tournament-style arenas. Then I considered for each season the number of NCAA wins the team achieved.
Note that I included ACC/Big East and ACC/Big 10 challenge games, Wooden Classic and Coaches v. Cancer games wherever they were played, as well as other pre-season or in-season tournaments that we played in with the exception of the Hawaii and Alaska tournaments, because while those were neutral sites, the gyms were so small that the atmospheres were nothing like an NCAA tournament game. So they didn’t fit into any of the three categories described above. We’ve also played pre-season tournament games at places like Kansas City’s Municipal Auditorium, in front of crowds of 7000 or 8000. I didn’t count those, as those aren’t NCAA tournament-like environments. I considered games played at the Greensboro Coliseum as being played at a neutral site/tournament-like arena, assuming the opponent was the type I described above, not someone like UNC-G or NC A&T. Madison Square Garden games were always considered as neutral site, tournament-style games regardless of whether they were NIT Pre-Season tournament games, made-for-TV games like the JJ game against Texas or the recent Gonzaga and Washington games, or games played as part of the St. John’s series.
So here are the numbers, then I’ll take a crack at analyzing them.
|
Neutral Tournament-Style |
Road On Campus |
Home |
NCAA Wins |
83-84 |
0 |
0 |
0 |
0 |
84-85 |
1 |
0 |
1 |
1 |
85-86 |
2 |
1 |
2 |
5 |
86-87 |
1 |
1 |
0 |
2 |
87-88 |
0 |
2 |
1 |
4 |
88-89 |
1 |
2 |
1 |
4 |
89-90 |
1 |
1 |
3 |
5 |
90-91 |
3 |
2 |
2 |
6 |
91-92 |
1 |
3 |
1 |
6 |
92-93 |
1 |
1 |
4 |
1 |
93-94 |
0 |
2 |
2 |
5 |
94-95 |
2 |
2 |
1 |
0 |
95-96 |
1 |
1 |
2 |
0 |
96-97 |
3 |
1 |
1 |
1 |
97-98 |
0 |
1 |
2 |
3 |
98-99 |
4 |
0 |
2 |
5 |
99-00 |
4 |
1 |
2 |
2 |
00-01 |
5 |
1 |
2 |
6 |
01-02 |
2 |
1 |
2 |
2 |
02-03 |
3 |
0 |
2 |
2 |
03-04 |
2 |
1 |
1 |
4 |
04-05 |
3 |
0 |
2 |
2 |
05-06 |
5 |
1 |
1 |
2 |
06-07 |
2 |
0 |
3 |
0 |
07-08 |
2 |
0 |
3 |
1 |
08-09 |
4 |
2 |
1 |
2 |
09-10 |
5 |
1 |
1 |
6 |
10-11 |
5 |
0 |
2 |
2 |
11-12 |
3 |
1 |
1 |
0 |
What do these numbers tell us? Well, first of all, I do see a shift in the number of neutral site tournament-style games we’ve played as opposed to on-campus road games, and it appears that shift began in either 1996-97 or 1998-99. Let’s call it ’96-’97. From 83-84 through 95-96 we played a total of 14 neutral site tournament style games in those 13 years, for an average of 1.07 per year. But during those same 13 years we played 18 on-campus road games, for an average of 1.38 per year, and 20 home games, for an average of 1.54 per year. The most neutral games we had in any of those 13 years was the three we played in 1990-91, and one of those was actually a quasi-road game against Georgetown at the old Capital Centre.
But starting in 1996-97 and running through 2011-12, we played a total of 52 neutral site tournament style games in those 16 years, for an average of 3.25 per year. This total includes a zero for 1997-98. In those same years we played only 11 on-campus road games, for an average of 0.69 per year. We played 28 home games against quality nonconference opponents in those 16 years, for an average of 1.75 per year.
So utilizing 1996-97 as a cutoff point and looking at our scheduling pre- and post-that season, the reality is that while we continued to schedule approximately the same number of home games against these types of opponents, we cut our on-campus road games by half while we more than tripled the number of neutral site, tournament-style games we played.
That’s the answer to question number one.
Question number two is much more difficult to answer: what effect has increasing our scheduling of neutral site, tournament-style games had on our actual tournament performance?
The first thing I looked at was our championship years of 1991, 1992, 2001 and 2010. In ’91 we played 3 neutrals and two on the road, and in ’92 it was only one neutral and three on the road. Not much to say there except that the great ’92 team was able to go the distance, including winning the last two in the 50,000 seat Metrodome, without the benefit of much tournament-arena experience, as even the one they did have was in Greensboro against St. John’s. But that was a veteran team and had the experience of the three neutrals from the prior season, for whatever that’s worth.
The most neutrals we’ve ever played in a season is five, which we’ve done four times, including two of the least three seasons. One of those four times was the championship year of 2001, and in that year we played only one road game of the sort I’m talking about in this analysis, at Temple. I counted the Stanford game at the Oakland Coliseum as a neutral, as it was not on campus and was at a tournament-style arena.
Likewise, in the championship year of 2010 we played five neutrals and only one road game, against Wisconsin at the Kohl Center, which seats over 17,000. Could’ve easily counted that one as a neutral in that it’s a tournament-like environment there, but didn’t because it’s on campus. On the other hand, I counted the Georgetown game at the Verizon Center as a neutral, as it’s big and not on campus.
There were only two other years that we played five neutrals. One was 2011, and that year we played no roadies (of the type I’m considering) at all. I counted the Singler game against Oregon at the Rose Garden as a neutral. It’s difficult to really consider this year in any kind of analysis, as Kyrie’s injury disfigured the entire season, including of course the way we played in the humbling loss to Arizona in Anaheim.
The other was 2006, where it was one roadie against the five neutrals, and yet we went out to LSU in the Georgia Dome. That was not a good night.
What about other years we’ve made the Final Four? What did the scheduling look like in those years? Well, a lot of them were in the late ‘80s and early ‘90s when we were scheduling more roadies and less neutral site games. 1988: no neutrals, 2 roadies. 1994: no neutrals, two roadies. But then there was 1986 where there were two neutrals and one roadie, and 1991 with 3 neutrals and two roadies. Not much to make of that era, really. Mixed bag.
But of more recent vintage, the 1999 Final Four team played four neutrals and no roadies. The 2004 Final Four team played two neutrals and one roadie as again I counted the Georgetown game at the MCI Center as a neutral, which I know can be argued.
Another way to look at the data is as follows: we’ve played five neutrals in four different seasons, and those seasons resulted in two championships, one sweet 16 loss, and one “incomplete” due to Kyrie’s injury.
Taking this data together, it appears that in many of the years in which we’ve reached the Final Four we have indeed utilized a scheduling formula that included more games at neutral site, tournament-type arenas, and less at quality opponents’ on-campus arenas.
But it’s not that clear cut. Why not? Because in a number of other years, we’ve utilized that formula, and had little or no success in the NCAA’s.
What about when we’ve played four neutrals? We’ve done that three times, in 1999, 2000, and 2009. We reached the final game in ’99 (which we lost to UConn due solely to the fact that I was in China and couldn’t watch the game), and lost in the Sweet 16 in 2000 and 2009 Not so good.
What about three neutrals? In 1997 we played three neutrals and one roadie, but got bounced in the second round by Providence. The program was on the rebound from the trauma of 1995, and clearly Duke was not a top tier team, but even if discounted, it still has to be counted.
In 2003 we played three neutrals and no roadies and again went out in the Sweet 16, this time to Kansas. Young, rebuilding team, but it counts. In 2005 it was three neutrals and no roadies and out in the Sweet 16 to Michigan State. And then even last year where we had the three neutrals (Michigan State, Washington, & Temple) and one roadie (Ohio State) and then came Lehigh.
So what that looks like to me is that even when we’ve picked up our neutral-site games to three or even four, we haven’t had any corresponding type of success in the NCAA tournament. It’s not until we’ve really extended the philosophy to play five such games that we’ve had greater success, but because we’ve only done that four times (two titles), it’s a pretty small sample size from which to draw many conclusions.
Another perspective on this: what about when we have played zero neutral-site tournament arena games? How have we done? Well, it’s only happened four times, including all the way back in 1984, but the answer is “quite well, thank you.” In ’84, which again was K’s first NCAA tournament team, we didn’t play any of these types of quality opponents anywhere – home, road, or neutral, and the young team lost in the first round. In 1988 we played no neutrals but made the Final Four. In 1994 we played none and made the final game. And in 1998 we played none and lost in the Elite Eight to Kentucky, which never should’ve happened.
So looking at this history, I’d say this. There’s no question that K’s scheduling philosophy changed about 15 years ago. But I don’t think the NCAA results support the position taken by some that scheduling more neutral-site, tournament-like arena games in the regular season and playing correspondingly less on-campus road games against quality teams leads directly to greater success in the tournament, with the possible exception of when we really ramp up the neutrals and play five of them in a single season. But really, the numbers are all over the map. Some of our championship and Final Four years we’ve played a lot of neutrals, in others not really, but we won big anyway. Some years where we’ve played a lot of neutrals, we’ve had great success in the tournament. Other years, not. In a number of years when we’ve played no neutrals at all, we’ve done quite well in the tournament.
One last thing I noticed though. Maybe it’s not where these games against these quality nonconference opponents are played. Maybe it’s the mere fact of playing them at all, no matter where, that leads demonstrably to NCAA success. Well, maybe so. The highest number of these types of games we’ve played was the eight in the national championship year of 2001. We played seven such games a number of times, including 1991 (championship), 2000 (Sweet 16), 2006 (Sweet 16), 2009 (Sweet 16), 2010 (championship) and 2011 (Sweet 16/Kyrie). That’s two times out of six with a national championship, and three out of seven championships when we’ve played either 7 or 8 such nonconference games. When we’ve played six such games, we lost in the second round once (1993) and went to the final game once (1999).
Note that in thinking about it this way – playing quality teams no matter where, the numbers are somewhat skewed because I haven’t included in my analysis games, for instance, in Hawaii and Alaska tournaments, where we’ve of course played many, many excellent teams, because I didn’t want to go back and recalculate all the totals for each year when I came up with this hypothesis just as I was about to post this. I'm also aware that it's possible that scheduled more of these types of games in years he knew he had a strong, championship-caliber team, a team that could handle a more demanding schedule, and if so, that weakens any cause-and-effect argument that one might make.
But keeping in mind that my original goal here was to try to determine if increasing the neutral-site, tournament-style arena games and scheduling less on-campus road and home games corresponds with increased NCAA Tournament success, all I can really say is that I don't think it does. Now maybe somebody out there smarter than I am can crunch these numbers in a more sophisticated way than I can and come up with different conclusions. I understand that there are other reasons to play these large arena neutral site games, including revenue enhancement, recruiting, and playing in front of larger numbers of alumni. But if – if – the sole reason to be playing these games at these sites is to improve our chances in the NCAA tournament, the numbers don’t appear to support it. Better idea: have the better team.