WEEK 2 · Sun, Sep 20, 3:25 PM CDT

Miami Dolphins @ San Francisco 49ers

12 qualified model plays · 7 modeled lines did not qualify · 7-5-0
← Back to Week 2
MODEL OVERVIEW

What the board is seeing

Qualified plays 12 Published sides that cleared the current model thresholds.
Avg model win 67.3% Average estimated win probability across this game's qualified plays.
Avg no-vig edge +15.6% Model probability minus the market's no-vig implied probability.
Avg model-implied EV +22.6% Average modeled return using the published side price.

Projection gap is the raw difference between PDR's projection and the sportsbook line. No-vig edge is the probability advantage after removing sportsbook hold. They answer different questions.

How the model works →
QUALIFIED PLAYS

Game prop board

Ranked by the frozen published model board.

#1 MODEL PLAY

De'Von Achane

Receptions
WIN
UNDER 4.5
fanduel -168
Projection 3.4
Proj. gap -1.1
Model win 78.3%
No-vig market 58.6%
No-vig edge +19.7%
Model-implied EV +24.9%
WHY THIS PLAY
Projection 3.4 is 1.1 below the 4.5 line.
Model win probability 78.3% vs 58.6% no-vig market probability creates a +19.7% edge.
Qualified with 2154 residual samples, 78.3% model win, +19.7% no-vig edge, and +24.9% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 3 receptions
#2 MODEL PLAY

Christian McCaffrey

Receiving yards
LOSS
UNDER 42.5
fanduel -114
Projection 33.2
Proj. gap -9.3
Model win 74.9%
No-vig market 50.0%
No-vig edge +24.9%
Model-implied EV +40.7%
WHY THIS PLAY
Projection 33.2 is 9.3 below the 42.5 line.
Model win probability 74.9% vs 50.0% no-vig market probability creates a +24.9% edge.
Qualified with 2154 residual samples, 74.9% model win, +24.9% no-vig edge, and +40.7% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 43 receiving yards
#3 MODEL PLAY

Christian McCaffrey

Rushing yards
LOSS
OVER 57.5
fanduel -114
Projection 83.1
Proj. gap +25.6
Model win 70.7%
No-vig market 50.0%
No-vig edge +20.7%
Model-implied EV +32.8%
WHY THIS PLAY
Projection 83.1 is 25.6 above the 57.5 line.
Model win probability 70.7% vs 50.0% no-vig market probability creates a +20.7% edge.
Qualified with 796 residual samples, 70.7% model win, +20.7% no-vig edge, and +32.8% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 796
Final: 23 rushing yards
#4 MODEL PLAY

Malik Willis

Rushing yards
WIN
UNDER 36.5
fanduel -114
Projection 29.7
Proj. gap -6.8
Model win 70.5%
No-vig market 50.0%
No-vig edge +20.5%
Model-implied EV +32.3%
WHY THIS PLAY
Projection 29.7 is 6.8 below the 36.5 line.
Model win probability 70.5% vs 50.0% no-vig market probability creates a +20.5% edge.
Qualified with 796 residual samples, 70.5% model win, +20.5% no-vig edge, and +32.3% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 796
Final: 6 rushing yards
#5 MODEL PLAY

Brock Purdy

Passing TDs
WIN
OVER 1.5
fanduel -188
Projection 2.3
Proj. gap +0.8
Model win 70.2%
No-vig market 61.0%
No-vig edge +9.1%
Model-implied EV +7.5%
WHY THIS PLAY
Projection 2.3 is 0.8 above the 1.5 line.
Model win probability 70.2% vs 61.0% no-vig market probability creates a +9.1% edge.
Qualified with 516 residual samples, 70.2% model win, +9.1% no-vig edge, and +7.5% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 516
Final: 2 passing TDs
#6 MODEL PLAY

George Kittle

Receptions
WIN
OVER 3.5
fanduel -154
Projection 5.2
Proj. gap +1.7
Model win 66.9%
No-vig market 56.7%
No-vig edge +10.1%
Model-implied EV +10.3%
WHY THIS PLAY
Projection 5.2 is 1.7 above the 3.5 line.
Model win probability 66.9% vs 56.7% no-vig market probability creates a +10.1% edge.
Qualified with 2154 residual samples, 66.9% model win, +10.1% no-vig edge, and +10.3% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 4 receptions
#7 MODEL PLAY

Deebo Samuel Sr.

Receiving yards
LOSS
OVER 32.5
fanduel -114
Projection 54.2
Proj. gap +21.7
Model win 65.5%
No-vig market 50.0%
No-vig edge +15.5%
Model-implied EV +22.9%
WHY THIS PLAY
Projection 54.2 is 21.7 above the 32.5 line.
Model win probability 65.5% vs 50.0% no-vig market probability creates a +15.5% edge.
Qualified with 2154 residual samples, 65.5% model win, +15.5% no-vig edge, and +22.9% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 31 receiving yards
#8 MODEL PLAY

Greg Dulcich

Receiving yards
LOSS
OVER 23.5
fanduel -114
Projection 44.7
Proj. gap +21.2
Model win 64.4%
No-vig market 50.0%
No-vig edge +14.4%
Model-implied EV +20.9%
WHY THIS PLAY
Projection 44.7 is 21.2 above the 23.5 line.
Model win probability 64.4% vs 50.0% no-vig market probability creates a +14.4% edge.
Qualified with 2154 residual samples, 64.4% model win, +14.4% no-vig edge, and +20.9% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 17 receiving yards
#9 MODEL PLAY

Christian McCaffrey

Receptions
WIN
UNDER 4.5
fanduel +112
Projection 4.5
Proj. gap -0.0
Model win 64.2%
No-vig market 44.1%
No-vig edge +20.0%
Model-implied EV +36.0%
WHY THIS PLAY
Projection 4.5 is 0.0 below the 4.5 line.
Model win probability 64.2% vs 44.1% no-vig market probability creates a +20.0% edge.
Qualified with 2154 residual samples, 64.2% model win, +20.0% no-vig edge, and +36.0% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 4 receptions
#10 MODEL PLAY

De'Von Achane

Rushing yards
WIN
OVER 54.5
fanduel -114
Projection 74.3
Proj. gap +19.8
Model win 61.8%
No-vig market 50.0%
No-vig edge +11.8%
Model-implied EV +16.0%
WHY THIS PLAY
Projection 74.3 is 19.8 above the 54.5 line.
Model win probability 61.8% vs 50.0% no-vig market probability creates a +11.8% edge.
Qualified with 796 residual samples, 61.8% model win, +11.8% no-vig edge, and +16.0% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 796
Final: 74 rushing yards
#11 MODEL PLAY

Malik Willis

Passing yards
LOSS
UNDER 184.5
fanduel -114
Projection 163.4
Proj. gap -21.1
Model win 61.0%
No-vig market 50.0%
No-vig edge +11.0%
Model-implied EV +14.6%
WHY THIS PLAY
Projection 163.4 is 21.1 below the 184.5 line.
Model win probability 61.0% vs 50.0% no-vig market probability creates a +11.0% edge.
Qualified with 516 residual samples, 61.0% model win, +11.0% no-vig edge, and +14.6% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 516
Final: 197 passing yards
#12 MODEL PLAY

George Kittle

Receiving yards
WIN
OVER 42.5
fanduel -114
Projection 60.3
Proj. gap +17.8
Model win 59.7%
No-vig market 50.0%
No-vig edge +9.7%
Model-implied EV +12.0%
WHY THIS PLAY
Projection 60.3 is 17.8 above the 42.5 line.
Model win probability 59.7% vs 50.0% no-vig market probability creates a +9.7% edge.
Qualified with 2154 residual samples, 59.7% model win, +9.7% no-vig edge, and +12.0% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 80 receiving yards
MODELED - DID NOT QUALIFY

Other lines PDR reviewed

Context only. These are not published recommendations and are not graded.

MODELED - DID NOT QUALIFY

Brock Purdy

Passing yards
Did not qualify
Model side: OVER 238.5
fanduel -114
Projection246.6
Proj. gap+8.1
Model win52.7%
No-vig market50.0%
No-vig edge+2.7%
Model-implied EV-1.0%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 516
MODELED - DID NOT QUALIFY

De'Von Achane

Receiving yards
Did not qualify
Model side: OVER 27.5
fanduel -114
Projection29.8
Proj. gap+2.3
Model win36.4%
No-vig market50.0%
No-vig edge-13.6%
Model-implied EV-31.8%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 2154
MODELED - DID NOT QUALIFY

Deebo Samuel Sr.

Receptions
Did not qualify
Model side: OVER 3.5
fanduel +142
Projection4.1
Proj. gap+0.6
Model win46.3%
No-vig market38.6%
No-vig edge+7.7%
Model-implied EV+12.0%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 2154
MODELED - DID NOT QUALIFY

Greg Dulcich

Receptions
Did not qualify
Model side: OVER 2.5
fanduel +124
Projection3.4
Proj. gap+0.9
Model win51.8%
No-vig market41.7%
No-vig edge+10.1%
Model-implied EV+16.0%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 2154
MODELED - DID NOT QUALIFY

Malik Willis

Passing TDs
Did not qualify
Model side: OVER 0.5
fanduel -162
Projection0.8
Proj. gap+0.3
Model win55.2%
No-vig market57.9%
No-vig edge-2.6%
Model-implied EV-10.7%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 516
MODELED - DID NOT QUALIFY

Mike Evans

Receptions
Did not qualify
Model side: OVER 3.5
fanduel -114
Projection4.2
Proj. gap+0.7
Model win48.2%
No-vig market50.0%
No-vig edge-1.8%
Model-implied EV-9.5%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 2154
MODELED - DID NOT QUALIFY

Mike Evans

Receiving yards
Did not qualify
Model side: OVER 47.5
fanduel -114
Projection55.5
Proj. gap+8.0
Model win44.2%
No-vig market50.0%
No-vig edge-5.8%
Model-implied EV-16.9%
Why it did not qualify: Model win probability did not clear the publication floor.
Historical residual sample: 2154
PDR is part of The Buffet. Turn your sports takes into challenges with friends at Buffet Bets →