Data scientist, physicist, and fantasy football champion

Week 13 DEF Evaluation

Last week (that is, Week 12) I used 6 models to predict which defenses would earn you the most fantasy points. Models A and B did very poorly that week despite having done well the rest of the season. Models D and E were always somewhat rubbish and continued to be so. And Models C and F were starting to catch up to the rest. In Week 12 model F was the best, so I trusted it to carry me through this week. Let’s see how these did this week:

Rank

Model.A

Model.B

Model.C

Model.D

Model.E

Model.F

Yahoo

DWC

True.Order

True.Score

1

TB

DEN

DEN

TB

DEN

DEN

DEN

DEN

DEN

16

2

DEN

TB

SEA

DEN

MIN

SEA

NE

SEA

CHI

15

3

SEA

SD

IND

IND

JAC

GB

SEA

NE

BAL

15

4

WAS

SEA

PHI

SD

HOU

WAS

BAL

BAL

TB

12

5

SD

WAS

WAS

SEA

SF

TB

SD

PHI

NE

12

6

IND

IND

TB

WAS

SD

PHI

GB

SD

IND

11

7

BAL

PHI

GB

PHI

NYJ

IND

PHI

ARI

DET

11

8

PHI

BAL

NE

JAC

MIA

NE

ARI

GB

KC

9

9

GB

GB

CAR

MIA

SEA

CIN

MIA

DAL

PIT

9

10

NE

NE

ARI

CAR

TB

JAC

BUF

SF

OAK

8

11

MIA

MIA

MIN

NE

OAK

SD

DAL

PIT

CIN

8

12

MIN

CHI

KC

DET

CAR

PIT

PIT

MIA

MIN

8

13

CAR

MIN

SD

BAL

KC

OAK

CIN

CHI

SEA

8

14

KC

KC

JAC

PIT

LA

BAL

MIN

OAK

GB

8

15

ARI

ARI

BAL

CHI

GB

CAR

NYG

NO

NYG

6

Model.A

Model.B

Model.C

Model.D

Model.E

Model.F

Yahoo

DWC

47

53

55

67

61

46

44

52

A huge rebound for models A and B after last week, and Model F continues to perform well. But the winner this week is actually Yahoo. As in Week 11 I forgot to write down the top FantasyPros experts predictions before Tuesday morning and now they’re lost forever (or cached somewhere). Yahoo and Defense Wins Championships both predicted Washington outside the top 15 and their 1 point was extremely underwhelming. They were the highest team from Model F available in my league so I picked them up. Had I listened to Model A and picked up TB I would have been rewarded. Instead I’m now out of the playoffs. Kaepernick’s -3 point game (including -1 per sack) had a little something do with that as well, but defense is a thing I’m actively trying to predict so I’m disappointed there too.

Models D and E are still garbage. I may finally take them out of my predictions.

In Week 14 I’ll continue to trust Model F since it’s been consistently good for the last 3 weeks now, but I’ll keep an eye out for any teams that Models A or B really like.

Week 14 DEF Prediction

Week 13 K Predictions