WEEK 3 · Sun, Sep 27, 3:05 PM CDT
Arizona Cardinals @ San Francisco 49ers
10 qualified model plays · 2 modeled lines did not qualify · 8-2-0
MODEL OVERVIEW
What the board is seeing
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.
WIN
OVER 44.5
fanduel -114
Projection
87.7
Proj. gap
+43.2
Model win
91.5%
No-vig market
50.0%
No-vig edge
+41.5%
Model-implied EV
+71.7%
WHY THIS PLAY
Projection 87.7 is 43.2 above the 44.5 line.
Model win probability 91.5% vs 50.0% no-vig market probability creates a +41.5% edge.
Qualified with 2154 residual samples, 91.5% model win, +41.5% no-vig edge, and +71.7% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 89 receiving yards
WIN
OVER 30.5
fanduel -114
Projection
62.1
Proj. gap
+31.6
Model win
79.9%
No-vig market
50.0%
No-vig edge
+29.9%
Model-implied EV
+50.0%
WHY THIS PLAY
Projection 62.1 is 31.6 above the 30.5 line.
Model win probability 79.9% vs 50.0% no-vig market probability creates a +29.9% edge.
Qualified with 2154 residual samples, 79.9% model win, +29.9% no-vig edge, and +50.0% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 40 receiving yards
LOSS
UNDER 7.5
fanduel -148
Projection
6.5
Proj. gap
-1.0
Model win
77.9%
No-vig market
55.9%
No-vig edge
+22.0%
Model-implied EV
+30.5%
WHY THIS PLAY
Projection 6.5 is 1.0 below the 7.5 line.
Model win probability 77.9% vs 55.9% no-vig market probability creates a +22.0% edge.
Qualified with 2154 residual samples, 77.9% model win, +22.0% no-vig edge, and +30.5% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 9 receptions
WIN
OVER 3.5
fanduel -148
Projection
5.6
Proj. gap
+2.1
Model win
74.0%
No-vig market
55.9%
No-vig edge
+18.1%
Model-implied EV
+24.0%
WHY THIS PLAY
Projection 5.6 is 2.1 above the 3.5 line.
Model win probability 74.0% vs 55.9% no-vig market probability creates a +18.1% edge.
Qualified with 2154 residual samples, 74.0% model win, +18.1% no-vig edge, and +24.0% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 11 receptions
WIN
OVER 55.5
fanduel -114
Projection
83.3
Proj. gap
+27.8
Model win
73.2%
No-vig market
50.0%
No-vig edge
+23.2%
Model-implied EV
+37.5%
WHY THIS PLAY
Projection 83.3 is 27.8 above the 55.5 line.
Model win probability 73.2% vs 50.0% no-vig market probability creates a +23.2% edge.
Qualified with 796 residual samples, 73.2% model win, +23.2% no-vig edge, and +37.5% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 796
Final: 75 rushing yards
WIN
OVER 1.5
fanduel -178
Projection
2.4
Proj. gap
+0.9
Model win
70.5%
No-vig market
59.8%
No-vig edge
+10.8%
Model-implied EV
+10.2%
WHY THIS PLAY
Projection 2.4 is 0.9 above the 1.5 line.
Model win probability 70.5% vs 59.8% no-vig market probability creates a +10.8% edge.
Qualified with 516 residual samples, 70.5% model win, +10.8% no-vig edge, and +10.2% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 516
Final: 4 passing TDs
LOSS
UNDER 67.5
fanduel -114
Projection
64.2
Proj. gap
-3.3
Model win
69.9%
No-vig market
50.0%
No-vig edge
+19.9%
Model-implied EV
+31.2%
WHY THIS PLAY
Projection 64.2 is 3.3 below the 67.5 line.
Model win probability 69.9% vs 50.0% no-vig market probability creates a +19.9% edge.
Qualified with 2154 residual samples, 69.9% model win, +19.9% no-vig edge, and +31.2% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 2154
Final: 75 receiving yards
WIN
OVER 2.5
fanduel +116
Projection
4.3
Proj. gap
+1.8
Model win
68.2%
No-vig market
43.3%
No-vig edge
+24.9%
Model-implied EV
+47.2%
WHY THIS PLAY
Projection 4.3 is 1.8 above the 2.5 line.
Model win probability 68.2% vs 43.3% no-vig market probability creates a +24.9% edge.
Qualified with 2154 residual samples, 68.2% model win, +24.9% no-vig edge, and +47.2% 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
WIN
OVER 1.5
fanduel +178
Projection
1.9
Proj. gap
+0.4
Model win
59.5%
No-vig market
33.6%
No-vig edge
+25.9%
Model-implied EV
+65.4%
WHY THIS PLAY
Projection 1.9 is 0.4 above the 1.5 line.
Model win probability 59.5% vs 33.6% no-vig market probability creates a +25.9% edge.
Qualified with 516 residual samples, 59.5% model win, +25.9% no-vig edge, and +65.4% 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
WIN
OVER 225.5
fanduel -114
Projection
252.1
Proj. gap
+26.6
Model win
59.1%
No-vig market
50.0%
No-vig edge
+9.1%
Model-implied EV
+11.0%
WHY THIS PLAY
Projection 252.1 is 26.6 above the 225.5 line.
Model win probability 59.1% vs 50.0% no-vig market probability creates a +9.1% edge.
Qualified with 516 residual samples, 59.1% model win, +9.1% no-vig edge, and +11.0% model-implied EV; publication floors are 500 samples, 57.5% win, +5.0% edge, and +5.0% EV.
Historical residual sample: 516
Final: 280 passing yards
MODELED - DID NOT QUALIFY
Other lines PDR reviewed
Context only. These are not published recommendations and are not graded.
Did not qualify
Model side: OVER 237.5
fanduel -114
Projection239.1
Proj. gap+1.6
Model win50.0%
No-vig market50.0%
No-vig edge+0.0%
Model-implied EV-6.1%
Why it did not qualify:
Model win probability did not clear the publication floor.
Historical residual sample: 516
Did not qualify
Model side: OVER 33.5
fanduel -114
Projection42.1
Proj. gap+8.6
Model win46.7%
No-vig market50.0%
No-vig edge-3.3%
Model-implied EV-12.3%
Why it did not qualify:
Model win probability did not clear the publication floor.
Historical residual sample: 796