Data scientist, physicist, and fantasy football champion

Week 9 DEF Predictions

Model F finally broke ahead of Yahoo’s experts, though only barely. It was the best model last year, and it continues to be so this year, but again, not statistically significantly better than Yahoo. I still like it, though, and will be using it for my predictions.

If things work out like they did last year, Model H might start pulling ahead soon. It’s the same as Model F, but uses only data from this year. About halfway through the season it should be easier to predict performance.

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 9 model summary:
 

RankACFGHI
1DETPHIPHIPHIJACPHI
2PHIHOUHOUHOUPHIHOU
3JACLARLARJACDETBUF
4HOUBUFJACLARSEAJAC
5SEAJACBUFARIBUFLAR
6INDDETARIBUFINDDET
7BALSEACARCARHOUTEN
8GBARISFTENNOSEA
9BUFCARTENSFGBCAR
10NOBALDETDETTENARI
11LARKCOAKSEAARIBAL
12DENTENNOOAKLARATL
13ARISFSEANOBALKC
14TENATLATLATLWASSF
15WASNYGDENKCDENNO
16CARNOKCDENNYJCIN
17NYJDENWASCINCAROAK
18CINOAKBALNYGSFDEN
19KCWASNYGBALATLNYG
20ATLCINCINGBDALGB
21OAKGBGBWASNYGWAS
22MIAINDNYJNYJOAKMIA
23DALTBTBDALKCNYJ
24TBDALINDTBCINDAL
25SFNYJDALMIAMIATB
26NYGMIAMIAINDTBIND

Conclusions


Unlike last week, Models F and H offer a lot of interesting streaming options this week:

  • LAR (at NYG): The Rams’ defense has been good, the Giants’ offense has been bad. This game has one of the lowest predicted scores of the week. I’d expect the Rams to sack Eli and keep him from throwing to his depleted WR core.
  • SF (v ARI): At home against Drew Stanton is a good bet. SF hasn’t been great for fantasy purposes, but they’ve gotten their share of sacks and interceptions this year. They should be able to effectively shut down Stanton
  • ARI (at SF): On the other side of the ball, ARI is a good start if Beathard is playing. If it’s Garoppolo, that’s a bit trickier. He’ll be on a new team, but he did alright last year playing for New England when Brady was out. I’d avoid ARI if he’s playing Week 9.

CAUTION:

  • CAR (v ATL): Atlanta’s offense has been but a shadow of its former self, but I’m always cautious when getting against them. The expected score is on the lower end (about 20 points), but if Devonta Freeman is out they might have to throw more.

Obviously if someone dropped HOU or JAC you should go get them, likewise if there’s anyone high on my lists this week that are oddly available in your league. Check first, then stream one of these guys and remember: save your bench for better things.

Week 8 FLEX rankings and accuracy scores

Week 9 QB Predictions