Defining Duke's Dominance

-Matt Cox


Sorry folks - it’s time to wake up from that Eggnog induced coma of bliss and re-enter the Matrix…

With the holiday break pause on college hoops about to resume, we now return to the dark, dark reality that has set in over the past few months - one in which the evilest of blue blood empires, ‘the Smurf Satans’ (as dubbed by my colleague Jim Root), has taken over supreme rule of the college basketball universe. What started as a ballooned hollow ball of hype has now become a planet-sized wrecking ball of destruction that wipes away anything and everything in its path. Duke’s demolition of Kentucky officially put the world on notice, and it was only a week later when the hot stove historical comparisons began to run rampant across social media.

This momentous buzz was epitomized by one of the most respected sports books in Las Vegas, the Westgate, releasing a hypothetical game spread for this year’s Duke team versus the 1992 Michigan Wolverines ‘Fab Five’ squad. Despite having only three games to evaluate the Duke juggernaut, the oddsmakers were convinced the 2019 Blue Devils would be favored by 7.5 points over the Fab Five on a neutral floor (if only time travel was a real thing - sigh).

Being the disciplined [and fraudulent] data scientist that I am / claim to be, I opted to let the hand play out and gather a larger sample size of game results before jumping to any similar outlandish conclusions…

*Fast forward six weeks*

Welp, the non-conference season is now in the books, and the Duke machine has showed no signs of slowing down…

Kentucky by 34… San Diego State by 26… Indiana by 21…

Duke is mercifully destroying perennial program powers as if they were Division 2 exhibition tune-ups. So far this season, they have outscored opponents by an average of 28 points a game and are on the verge of cracking’s number 1 overall ranking for both offense and defense - for context, this feat has not been accomplished since 2002 when the blue-tinted Durham Devils reached kenpom’s offensive and defensive efficiency summit (Indiana fans should fondly remember what derailed that freight train).


Since I had nothing better to do this holiday break, I decided to dig into some data to contextualize just how dominant Duke has been, relative to some of the historically great teams of the past decade. While I realize I’m likely excluding teams prior to 2008 who deserve to be considered - that aforementioned 2002 Duke team is an obvious omission - I hand picked eight teams from the past 10 seasons that, in my opinion, have stood head and shoulders above the rest.

Below is a table summary of the non-conference resumes for all nine teams (8 historical teams + 2019 Duke), with this year’s Duke team shown in the bottom row:

Data above represents non-conference games only, including record, offensive and defensive points per possession (‘PPP’) and overall strength of schedule (per

Upon a quick scan of the table above, no one team distinguishes itself from the rest of the pack, but four teams do appear to emerge as preliminary frontrunners:

  • Unsurprisingly, 1) the current 2019 Duke squad makes the cut due to an impressive 0.36 net efficiency margin against the 2nd toughest non-conference schedule of the nine teams shown above

  • The national champion 2) 2010 Duke team was also a force to be reckoned the first half of the season, sporting the 2nd highest efficiency margin of the nine teams above (0.38) and a formidable top-100 non-conference strength of schedule

  • 3) 2015 Kentucky used their Monstar-esque length to erase shots at an historic rate and Big Blue nation posted the highest net efficiency margin against the 4th toughest non-conference schedule of the teams identified

  • 4) 2015 Wisconsin (who would eventually knock off that 2015 UK team in the Final Four) posted the 5th highest net efficiency margin against a brutal non-conference slate that ranked in the top-30 nationally

That high-level efficiency overview acts as a nice filter, but an additional layer of analysis is needed to determine if one of those four juggernauts can truly be dubbed “the most dominant team of the past decade”. To help sort through some of this noise, we overlaid a handicapping-based comparable assessment. By using historical against-the-spread data, specifically closing game lines and ultimate game outcomes, we can extrapolate what bookmakers’ expectations of each team were and how each team performed relative to those expectations.

In essence, we’re inferring what the smartest guys in the room would say - that is, the Vegas oddsmakers - if we asked them for their opinion on this debate.

Below is another table summary of the non-conference results for all nine teams (8 historical teams+ 2019 Duke), this time from a handicapping lens:

Data above represents non-conference games only, including average margin of victory, average spread value, average margin of cover, as well as each team’s ATS record and win percentage (per and Highlighted cells indicate max, highest or ‘best’ value for each column.

Well whaddya know - those same four teams appear to rise to the top again...

While the current year Duke (Duke ‘19), Duke ‘10 and Kentucky ‘15 teams each blew away their non-conference foes by an average of ~27 points a game, the more methodically paced 2015 Badgers consistently slow-cooked their opponents as well and surpassed Vegas’ expectations in 10 of their 13 non-conference tilts.

This macro against-the-spread analysis confirms only a few teams deserve to be mentioned in the same breathe as the current Zion Williamson, RJ Barrett, Cam Reddish and Tre Jones led Duke team. But, the variance in strength of schedules across teams still clouds our interpretation of the data and prevents us from anointing the 2018-19 Duke team as THE most dominant of them all.

One way to cut through this complexity is to compare individual non-conference games of this year’s Duke team with the other eight historical teams. We searched through all of the non-conference games of all eight historical teams to find close matches with Duke’s non-conference matchups so far this season. Below, there are three sets of examples which show the following:

  • One of Duke’s non-conference games from this year (first row)

  • Two other non-conference games featuring one of the historical eight teams and a comparable opponent (second and third rows)

The ‘Comp 1’ and ‘Comp 2’ examples are similar in both the caliber of opponent (defined by opponent overall rank in the “Opponent KP Rank” column) and the location of where the game was played (shown in the “Game Location” column). This enables a more ‘apples-to-apples’ comparison between Duke and the other eight historically great teams, in terms of oddsmakers’ pregame expectations and actual game results.

In this first set below, we identified two precedents for Duke’s most recent bout with Texas Tech at Madison Square Garden. Despite converting just 3 of 20 from behind the arc and coughing up the rock a whopping 19 times, the Devils still pulled away from a fringe top-10 Texas Tech team on a neutral floor to win by 11.

As impressive as Duke’s performance was last week, this chart should make you appreciate just how surgical that 2015 Wisconsin team was. Against a comparable opponent (Oklahoma) on a neutral floor, the oddsmakers gave OU a fairly large cushion of 8.5 points, only to watch the Badge carve up the Sooners en route to a 13 point win (and cover).

The second and third sets below reinforce the same narrative about Duke as stated above. Not only is Vegas setting higher lines against comparable competition, but the Blue Devils are smashing those lofty expectations - just refer to the fairly comfortable covers against Yale (by 9) and Indiana (by 5.5):


To sum all of this up in a metaphor, think of Duke and the other eight historical teams as Olympic high jumpers and Las Vegas oddsmakers as their respective trainers. Each trainer sets the bar based on how high they think their athlete will jump and then adjusts that bar based on whether or not the athlete clears the bar and by how much they clear it.

Of the nine athletes competing in this hypothetical high jump competition, I’d put my money on 2019 Duke, 2015 Kentucky and 2015 Wisconsin to take home the gold, silver and bronze medals in some order, with 2010 Duke just missing the podium. When factoring in non-conference strength of schedule and comparable historical game results, these teams faced the highest expectations and surpassed those expectations more frequently and by a larger margin than the other dominant teams of the past decade.

But as far as crowning this year’s Duke squad THE most dominant of that ‘Fantastic 4’, the data simply doesn’t make a convincing enough argument to award them them the gold medal with a high degree of certainty.