I am going to try to identify common characteristics of players who have killed Duke's defense this season, and then see if any realistic NCAA Tournament opponents have this type of player. Alternatively, we could discover that a certain "type" of Duke killer doesn't exist, and the player who tears up Duke's defense to end Duke's season could be practically anyone.
Below are the players who have had high usage (24% or greater), high efficiency (120 or greater) games against Duke this season, while playing at least 20 minutes. Positions are designated based on kenpom depth charts from the past five games, which determine a player's position based on his stats relative to his teammates' stats. Bolded players are those who performed significantly better against Duke than their season averages.
Frank Mason, Kansas
5-11, 190, PG
Against Duke: 25%, 129
Season: 26%, 124
Xavier Rathan-Mayes, @Florida State
6-4, 208, PG
Against Duke: 27%, 140
Season: 21%, 112
Dennis Smith, NC State
6-3, 190, PG
Against Duke: 43%, 121
Season: 28%, 110
John Collins, @Wake Forest
6-10, 235, C
Against Duke: 38%, 127
Season: 30%, 124
Bryant Crawford, @Wake Forest
6-3, 200, PG
Against Duke: 32%, 134
Season: 26%, 116
Michael Young, Pittsburgh
6-9, 235, C
Against Duke: 28%, 133
Season: 30%, 111
Justin Jackson, North Carolina
6-8, 210, SF
Against Duke: 25%, 122
Season: 24%, 120
Shelton Mitchell, Clemson
6-3, 195, PG
Against Duke: 27%, 140
Season: 22%, 110
John Collins, @Wake Forest
6-10, 235, C
Against Duke: 33%, 145
Season: 30%, 124
Bryant Crawford, @Wake Forest
6-3, 200, PG
Against Duke: 25%, 130
Season: 26%, 116
John Gillon, @Syracuse
6-0, 178, PG
Against Duke: 26%, 169
Season: 21%, 117
Bruce Brown, @Miami
6-5, 190, SG/PG
Against Duke: 29%, 139
Season: 22%, 111
Dwayne Bacon, Florida State
6-7, 221, SF
Against Duke: 24%, 142
Season: 28%, 108
Bonzie Colson, Notre Dame
6-5, 225, C
Against Duke: 31%, 132
Season: 26%, 122
Of the 12 players who both killed Duke and performed significantly above their normal level against Duke, 7 are primarily PGs for their respective teams, 4 are Cs and 1 is a SF. I am counting Miami's Brown as primarily a PG to fit this narrative. I am counting Collins and Crawford twice each because they each killed Duke twice. I will refer to these 12 players as the "Duke Killers."
At first look, it seems we may legitimately have a "Matt Jones Effect" on our hands, which pains me to admit. Jones generally doesn't guard opposing PGs or Cs without a switch, because that would leave Duke's bigger players like Amile Jefferson and Harry Giles to guard guys who are too quick for them, or it would leave Duke's smaller/daintier players like Grayson Allen and Frank Jackson to guard guys who are too big for them (I am using "daintier" because Allen actually isn't significantly smaller than Jones).
Jones generally appears capable of guarding anyone in between PG and C. Apparently, he and Jayson Tatum have been doing a decent job of preventing the "in between" players from killing Duke above their normal levels. If Jones has been assigned to opposing PGs or Cs, I would assume the results have not been great, which would also lead to the list of players above. From this perspective, it seems that the eye test is being validated.
I will now move on to common characteristics of the Duke Killers, besides their positions of PG or C. My assumption is that free throw attempt rate will stand out, as it indicates a player who is either quick enough to break down a defense and get to the rim frequently or big enough to get in good position around the rim frequently. This would presumably be effective against Duke's usual defensive scheme, which was 7th in the nation in preventing 3pt attempts this season and 4th in 3pt% allowed, but 149th in 2pt% allowed.
Below are the free throw attempt rates of the Duke Killers in conference games, along with their conference ranking in the stat. All of the Duke Killers are from the ACC.
Rathan-Mayes: 41%, #15
Smith: 50%, #6
Collins: 54%, #4
Crawford: 46%, #11
Young: 41%, #18
Mitchell: 41%, #16
Gillon: 47%, #9
Brown: 41%, #19
Bacon: 24%, NR
Colson: 42%, #14
Bacon is the odd man out here, both in position (SF) and in free throw attempt rate. All other Duke Killers are PGs or Cs, and have top 20 free throw attempt rates in the ACC. If we consider Bacon a one-off fluke, the narrative and becomes clear: If Duke loses another game this season, we have good reason to believe the main reason will be a PG who can penetrate Duke's no-3s defense or, in a less likely event, a C whom Duke's big men cannot rebound with or otherwise handle inside.
We knew this already, though. This should set our minds at ease that we are right in how we envision Duke ultimately being killed in the NCAA Tournament. I may follow up by scouring Duke's region for these sorts of players, but that should be the easy part. I'm not sure any other common characteristic would be helpful to know about, since I suspect they would be things like having a high 2pt% or 3pt%, which would simply make the player an efficient player in general.