SAT 19:00 ET|BOBBY DODD STADIUMCONFTENSIONVACATEDTALENT GAP15|O/U57.5COLLISION75
AP · FPI
TEAM
PTS
2025
TAL
OFF
DEF
RET
PACE
SOS ▾
IMP
ML
SPREAD
AWAY
33
2-9
64
9
80
38%
80
47
30.0
—
-2.5
HOME
56
3-7
79
46
29
99%
31
44
27.5
—
+2.5
IMP = IMPLIED TEAM TOTAL FROM SPREAD + O/U · OFF AND DEF ARE SCOUT TEAM GRADES AS FBS PERCENTILES · EVERY MEASURED FIGURE IS 2020CONSENSUS · CONSENSUS
FINALDUKE33—GT56
CLOSED DUKE -2.5TOTAL 57.5GT COVERED · 89 POINTS · OVER
DUKE
DERIVED · PLAY BY PLAY
GT
OFFICIAL FIGURES ARE THE BOX SCORE AS SCORED. DERIVED FIGURES COME FROM PLAY BY PLAY WITH GARBAGE TIME REMOVED, SO THEY COUNT A DIFFERENT SET OF PLAYS AND WILL NOT TIE OUT AGAINST THE OFFICIAL COLUMN. A BRIGHTENED SIDE LEADS THAT ROW BY THE DIRECTION IN ITS TOOLTIP; TIME OF POSSESSION IS UNMARKED BECAUSE IT IS A TENDENCY, NOT A VIRTUE. SEASON AVERAGES ARE REGULAR SEASON ONLY. MARKET FIGURES ARE BOOK CONSENSUS, NEVER MODEL OUTPUT.
DATA BASISNOTHING ON THIS TAB IS MEASURED YET · 2020 PRESEASON · NO GAMES PLAYED
MEASURED0 OF 12 GAMES
Real 2026 data. It accumulates weekly and replaces carried figures metric by metric, on the schedules below.
CARRIED14 ROOMS
What each roster did last season, applied player by player. Transfers name the school.
PROGRAM PRIOR2 STAFFS
The programme's own prior identity — or, where the staff is new, the coach's career fingerprint with its sample.
PRESEASON FACT6 PER TEAM
Talent, recruiting, vacated usage, portal and coach status. Preseason inputs.
ABSENT2 CLASSES
Current season luck and situational splits. Shown absent with the reason, never estimated.
FLIPS TO MEASUREDABSENTNOT MEASURED. This asks how many weeks into a season each number becomes trustworthy, which is a within-season question. The back-test measures relationships between one season and the next, so it cannot answer it.
MARKET LINE VS TEAM PROFILEGAME LEVEL · WEEK 13 FIELD n=44
MARKET
TOTAL57.5
SPREADDUKE -2.5
IMPLIED — DUKE30.0
IMPLIED — GT27.5
SCORING ENVIRONMENT
35FBS PCTL
BOTH OFF × BOTH DEF
DISAGREEMENT14 SLOTSThe two rankings disagree by 14 positions on a 44-game board. Both figures are stated; the site does not resolve which is right.
PACE COLLISION
FAST × SLOW
CEILING CHARACTERIDENTITY · NOT QUALITY
IS THE RECORD REAL2020 · FULL SEASON
DUKERECORD MATCHES
2-9ACTUALvs2.9-8.1FROM SCORING
GAP TO SCORING PROFILE-0.9 W
ONE SCORE RECORD0-1
POINT MARGIN / GAME-13.3
DUKE won 2 of 11 against a scoring profile describing 2.9. Both are facts about the same season; the market prices the first.
GTRECORD MATCHES
3-7ACTUALvs2.6-7.4FROM SCORING
GAP TO SCORING PROFILE+0.4 W
ONE SCORE RECORD1-0
POINT MARGIN / GAME-12.9
GEORGIA TECH won 3 of 10 against a scoring profile describing 2.6. Both are facts about the same season; the market prices the first.
TEAM LEDGERDUKEGEORGIA TECH2020 PRESEASON · NO GAMES PLAYED
VACATED USAGE — AND DID THE ROSTER REFILL ITOPEN VOLUME IS A FACT, NOT AN ASSIGNMENT
VACATED USAGE BY ROOMLEFT OPEN, LARGEST FIRST
QB0%Quentin HarrisQB · SHARE NOT MEASURED
RB0%Elijah DeveauxRB · SHARE NOT MEASUREDROTATION1 FURTHER DEPARTURE
WR/TE0%Scott BraceyWR · SHARE NOT MEASUREDROTATION2 FURTHER DEPARTURES
OL0%NOBODY LEFTTHE WHOLE ROOM RETURNS
DL0%Tahj RiceDT · SHARE NOT MEASUREDROTATION3 FURTHER DEPARTURES
LB0%Xander GagnonLB · SHARE NOT MEASUREDROTATION3 FURTHER DEPARTURES
DB0%Dylan SingletonS · SHARE NOT MEASUREDROTATION3 FURTHER DEPARTURES
SHARE OF LAST SEASON'S ROOM PARTICIPATION THAT DEPARTED, LARGEST FIRST · OPEN VOLUME IS A FACT, NOT AN ASSIGNMENT
PORTAL & CLASS INTAKE — WHO CAME INHEADCOUNT, NOT A PROJECTION
QBNO STARTING ROLE OPENEDINTACT
RBNO STARTING ROLE OPENEDINTACT
WR/TEJontavis Robertson · SIGNEE ★3Malik Bowen-Sims · SIGNEE ★3Nicky Dalmolin · SIGNEE ★3INTACT
OLNO DEPARTURE TO REFILLINTACT
DLAeneas Peebles · SIGNEE ★3Michael Reese · SIGNEE ★3INTACT
LBDorian Mausi Jr. · SIGNEE ★3Christian Hood · SIGNEE ★3Ryan Smith · SIGNEE ★3INTACT
DBIsaiah Fisher-Smith · SIGNEE ★3INTACT
PORTAL AND SIGNING-CLASS ARRIVALS AT EACH ROOM THAT LOST VOLUME · A HEADCOUNT, NEVER A USAGE PROJECTION
CARRIED GRADES & TEAM FACTS — THIS ROSTER, TODAYCARRIED FROM 2025, RE-HOMED PLAYER BY PLAYER · DUKE
CARRIED COMPOSITE
0.00 PCTL
ROOM
GRADE
FBS PCTL
OPP-ADJ
TAL
RET
N
OBS
QB———47CARRIED FROM 2019 · NOTHING GRADED YET
RB———898CARRIED FROM 2019 · NOTHING GRADED YET
WR/TE————25CARRIED FROM 2019 · NOTHING GRADED YET
OL————17CARRIED FROM 2019 · NOTHING GRADED YET
DL————18CARRIED FROM 2019 · NOTHING GRADED YET
LB————7CARRIED FROM 2019 · NOTHING GRADED YET
DB————18CARRIED FROM 2019 · NOTHING GRADED YET
TEAM FACTS — FULLY KNOWN NOWPRESEASON FACT
TALENT COMPOSITE64th64
BLUE-CHIP RATIO6%54
RECRUITING CLASS#61 · 0/18 BC55
CLASS RANK TREND+2 · 3 CYCLES54
COACH STATUSRETURNING · 5th YR—
DEV TRANSITION MIX23%15
EVERY ROW HERE IS KNOWABLE TODAY — NONE OF IT WAITS ON A GAME
COACH PRIOR — WHO APPLIES, AND WHAT A CHANGE MOVESBASIS · PROGRAM PRIOR
GTGEOFF COLLINS2TH SEASONNO CAREER RECORD
The programme's own prior season applies, under the same returning staff.
HOW MUCH IDENTITY MOVES WITH A RETURNING STAFF
|Δ PACE|1.37sn=929
|Δ EXPLOSIVENESS|6.72%n=931
|Δ HAVOC|2.10%n=929
|Δ PASS RATE|4.28%n=929
|Δ PASS RATE OVER EXPECTED|4.07%n=311
MEDIAN ABSOLUTE SEASON-TO-SEASON SHIFT · 2016-2026 · MAGNITUDE, NOT DIRECTION — A NEW STAFF MOVES PASS RATE AND PACE BARELY MORE THAN A RETURNING ONE
DUKEDAVID CUTCLIFFE5TH SEASONNO CAREER RECORD
The programme's own prior season applies, under the same returning staff.
HOW MUCH IDENTITY MOVES WITH A RETURNING STAFF
|Δ PACE|1.37sn=929
|Δ EXPLOSIVENESS|6.72%n=931
|Δ HAVOC|2.10%n=929
|Δ PASS RATE|4.28%n=929
|Δ PASS RATE OVER EXPECTED|4.07%n=311
MEDIAN ABSOLUTE SEASON-TO-SEASON SHIFT · 2016-2026 · MAGNITUDE, NOT DIRECTION — A NEW STAFF MOVES PASS RATE AND PACE BARELY MORE THAN A RETURNING ONE
POSITION GROUPS — WHERE THE VOLUME LIVESDUKE OFFENSEGT OFFENSERETURNINGDEPARTEDCOLLAPSE ALL
▾QBDUKE OFFENSEFULL DEFENSEGEORGIA TECH DEFENSE DUKE QUARTERBACKSn=796% VACATED
QUARTERBACKS · PLAYED THIS SEASON FIRST
OVR
EXP
PT%
TAL
★REC
vsCOH
PARTICIPATION SHARE IS OF THIS ROOM · THESE SIX COLUMNS DESCRIBE THE ROSTER · UNIT-AGAINST-UNIT AND USAGE DETAIL IS ON THE MATCHUP TAB
GT DEFENSE — GROUP GRADES3 ROOMS · 61 PLAYERSRETURNING SHARE NOT MEASURED
DEFENSIVE ROOMS · GROUP GRADES
GRD
RET
BLU
THE QUARTERBACK PLAYS THE WHOLE SIDE — SHOWN AS ITS THREE ROOM GRADES, NOT MAN FOR MAN · PER PLAYER DETAIL LIVES IN THE OTHER DRAWERS
▾OLDUKE OFFENSEFRONT-7GEORGIA TECH DEFENSE DUKE OFFENSIVE LINEn=17RETURNING SHARE NOT MEASURED
FIVE-MAN LINE · PLAYED THIS SEASON FIRST
OVR
EXP
PT%
TAL
★REC
vsCOH
PARTICIPATION SHARE IS OF THIS ROOM · THESE SIX COLUMNS DESCRIBE THE ROSTER · UNIT-AGAINST-UNIT AND USAGE DETAIL IS ON THE MATCHUP TAB
GT FRONT-7n=37 · 2 ROOMSRETURNING SHARE NOT MEASURED
PASS RUSH · PLAYED THIS SEASON FIRST
OVR
EXP
PT%
TAL
★REC
vsCOH
THE DEFENDERS THIS ROOM FACES, BY GRADE · THESE SIX COLUMNS DESCRIBE THE ROSTER · PASS RUSH AND COVERAGE DETAIL LIVES ON THE MATCHUP TAB
▾RBDUKE OFFENSEFRONT-7GEORGIA TECH DEFENSE DUKE RUNNING BACKSn=811% VACATED
BACKFIELD · PLAYED THIS SEASON FIRST
OVR
EXP
PT%
TAL
★REC
vsCOH
PARTICIPATION SHARE IS OF THIS ROOM · THESE SIX COLUMNS DESCRIBE THE ROSTER · UNIT-AGAINST-UNIT AND USAGE DETAIL IS ON THE MATCHUP TAB
GT FRONT-72 ROOMS · 37 PLAYERSRETURNING SHARE NOT MEASURED
ROOMS IN THE COMPOSITE
WT
GRD
RET
BLU
EACH ROOM ON ITS OWN GRADE, NOT LISTED MAN FOR MAN · THE WEIGHTS DEFINE WHO PLAYS WHOM
▾WR/TEDUKE OFFENSESECONDARYGEORGIA TECH DEFENSE DUKE PASS CATCHERSn=25RETURNING SHARE NOT MEASURED
PASS CATCHERS · PLAYED THIS SEASON FIRST
OVR
EXP
PT%
TAL
★REC
vsCOH
PARTICIPATION SHARE IS OF THIS ROOM · THESE SIX COLUMNS DESCRIBE THE ROSTER · UNIT-AGAINST-UNIT AND USAGE DETAIL IS ON THE MATCHUP TAB
GT SECONDARYn=24 · 1 ROOMRETURNING SHARE NOT MEASURED
COVERAGE · PLAYED THIS SEASON FIRST
OVR
EXP
PT%
TAL
★REC
vsCOH
THE DEFENDERS THIS ROOM FACES, BY GRADE · THESE SIX COLUMNS DESCRIBE THE ROSTER · PASS RUSH AND COVERAGE DETAIL LIVES ON THE MATCHUP TAB
RETURNINGPORTAL — PRIOR SCHOOL AND PORTAL STARS NAMEDLAST YR ≈ — RATES FROM HIS LAST MEASURED SEASON, NOT THE PROJECTED ROLEDEPARTEDVS COH = PRODUCTION vs OWN RECRUITING COHORT MEDIAN · CURVE = AHEAD / ON / BEHIND AT ±8 · PT% IS HIS SHARE OF HIS OWN ROOM · UNIT-AGAINST-UNIT AND USAGE DETAIL IS ON THE MATCHUP TAB
TEAM QUALITYHOUSE GRADE + REFERENCES
METRIC
DUKE
FBS PCTL 0 ── 100
GT
TALENT − PRODUCTION ▾+29
+36
SCOUT TEAM GRADE DECOMPOSES INTO OFFENSE AND DEFENSE, EACH BUILT FROM THE POSITION ROOM GRADES ABOVE · ROOM WEIGHTS ARE MEASURED AGAINST UNIT EPA, NOT ASSUMED · SPECIAL TEAMS ARE NOT IN IT YET
TEAM IDENTITYTENDENCIES THAT SET VOLUME
METRIC
DUKE
FBS PCTL 0 ── 100
GT
RUN RATE — STANDARD DNS56.9%
61.8%
PASS RATE — PASSING DNS73.1%
55.1%
GARBAGE TIME SHARE0.1%
0.5%
EVERY ROW HERE IS COUNTED FROM PLAY BY PLAY · NONE OF IT IS A QUALITY RATING — FURTHER RIGHT IS MORE OF THE THING, NOT BETTER · SCHEME DETAIL LIVES ON THE MATCHUP TAB
EVERY PERCENTILE READS HIGH-IS-GOOD · CELL SHADE = FBS PERCENTILE OF THE VALUE · DOTS PLOT BOTH TEAMS ON ONE 0–100 TRACK (GOLD = DUKE, ORANGE = GT) · MEASURED FIGURES ARE 2020, GARBAGE-TIME FILTERED
≈ MARKS AN ESTIMATED VALUE · A CARRIED FIGURE IS WHAT A PLAYER DID LAST SEASON · PORTAL MARKS PARTICIPATION EARNED AT THE NAMED SCHOOL
VACATED VOLUME IS STATED AS OPEN AND NEVER ASSIGNED TO A SUCCESSOR · SPREAD, TOTAL, ML AND IMPLIED TOTALS ARE CONSENSUS MARKET FIGURES · NO PROJECTIONS ON THIS PAGE