Data scientist, physicist, and fantasy football champion

Week 11 DEF Predictions

Another successful week for me personally, defense-wise (I picked NE), though I still lost (I also picked Aaron Jones)

Model F is still a little behind Yahoo’s experts again last week, but there’s very little daylight between any of the models. The odds have a lot of teams winning by a TD or more and predict some pretty low scores this week, so it should be a pretty good week for defenses.

Let’s see some predictions for this week:

ModelTypeScoreOppScoreOppHomeAwayDataYearsPointSystem
ALMXXXX2017IOUH/Standard
CLMXXXX2015-2017IOUH/Standard
FLMXXXX2015-2017TruScor
GGibbs SamplerXXXX2015-2017TruScor
HLMXXXX2017TruScor
IMany LMsXXXX2015-2017IOUH/Standard

 

Model A:

.Model A-1.png


Model A parameters:

The team that’s playing
Who their opponent is
Their predicted score
Their opponent’s predicted score
Whether the team is home or away
Model A Notes:

Uses only data from 2017
Standard points
Did well in 2016, but didn’t stand out

Model C:

.Model C-1.png


Model C parameters:

The team that’s playing
Who their opponent is
Their predicted score
Their opponent’s predicted score
Whether the team is home or away
Model C Notes:

Uses data from 2015-2017
Standard scoring
Did well in 2016, but didn’t stand out

Model F:

.Model F-1.png


Model F parameters:

The team that’s playing
Who their opponent is
Their predicted score
Their opponent’s predicted score
Whether the team is home or away
Model F Notes:

Uses data from 2015-2017
Truscor ignores defense TDs but weights interceptions and fumbles to try to account for it
I really liked this one in 2016. It was marginally better than A and C

Model G:

.Model G-1.png


Model G parameters:

The team that’s playing
Who their opponent is
Their predicted score
Their opponent’s predicted score
Whether the team is home or away
Model G Notes:

Uses data from 2015-2017
Truscor ignores defense TDs but weights interceptions and fumbles to try to account for it
Uses Bayesian statistics (Gibbs sampler) to create a model
Would have done very well in 2016 based on simulation

Model H:

.Model H-1.png


Model H parameters:

The team that’s playing
Who their opponent is
Their predicted score
Their opponent’s predicted score
Whether the team is home or away
Model H Notes:

Uses only data from 2017
Uses TruScor points to reduce the effect of TDs

Model I:

.Model I-1.png


Model I parameters may include:

The team that’s playing
Who their opponent is
Their predicted score
Their opponent’s predicted score
Whether the team is home or away
Model I Notes:

Data from 2015-2017
This model tries to predict all of the components of scoring separately (interceptions, sacks, etc.). Each component comes from a linear model containing only the significant terms (p < 0.05) from the list of possible terms above.

Week 11 model summary:
 

RankACFGHI
1JACJACJACJACJACJAC
2DETKCKCKCBALKC
3BALARIBALPITARIPIT
4NOBALARIDENDETBAL
5CHIPITDENBALCINDEN
6CINDENPITARIDENARI
7ARIPHICINCINNODET
8DENDETPHILACGBPHI
9SEACINHOUMIACHICIN
10TBMINLACHOUPITLAC
11PITHOUMIAPHILACMIA
12DALLACGBGBKCMIN
13HOUMIATBNODALTB
14MIATBDETSEASEANO
15GBNOLARDETMIASEA
16LACSEANOMINTBGB
17KCLARSEATBPHIHOU
18PHICLEMINLARMINCLE
19CLEGBBUFCLEHOUCHI
20WASBUFCLEBUFLARATL
21LARATLATLNECLEBUF
22MINCHINECHIWASLAR
23NENETENATLBUFNE
24BUFWASCHIDALNEDAL
25ATLDALDALTENATLTEN
26NYGNYGWASNYGTENNYG
27TENTENNYGWASNYGWAS
28OAKOAKOAKOAKOAKOAK

Conclusions

The reason why there’s so little difference between the models is because they all keep predicting the same top-6 or top-7 defenses as each other. This week it’s JAC, KC, ARI, BAL, DEN, PIT, and CIN. As with last week, most of these are taken, but here are some streaming options:

  • ARI (at HOU): They haven’t been great this year, but you can expect good things from a defense going against Tom Savage.
  • CIN (at DEN): They’re actually the underdogs this week, but neither team is expected to do much scoring. DEN has given up the most points to defenses this year. I’m aiming for Cincinatti this week.
  • MIA (v. TB): Going against a backup QB is usually worth a little bit. Miami has been good this year.
  • LAC (v. BUF): I was wrong about them last week (I thought they wouldn’t be able to contain the Jets. It’s a weird year), so I’m not making that mistake again. Buffalo allows a lot of sacks, though they have been scoring alright. I’m not excited about them, but you could do worse.

CAUTION:

  • GB (v. BAL): Green Bay has been surprisingly playable this year, but with the huge downgrade of their offense I’m wary to play their defense. Baltimore isn’t a high-scoring offense, but have only given up middle-of-the-road points. I’d still play GB this week if I had to, but I’d be willing to use DET, SEA, MIN, NO, or BUF first, and they’re all lower in my rankings.

If JAC, KC, BAL, PIT, PHI, or HOU are available, obviously go get them first. If not, I join you to stream with me. It’s worked out OK for me so far this year.

Week 11 K Predictions

Week 10 QB Results