Data scientist, physicist, and fantasy football champion

Week 9 QB Predictions

Model H pulled way ahead last week, so that’s definitely the go-to model for now. It uses only 2017 data to predict. I think something like this happened wih my models last year, too; around this time (halfway through the season) the most accurate model became the one using only 2017 data. This makes me think that there’s a Bayesian model out there that uses previous years as a weak prior, but there’s no way I’m coding that up any time soon.

Last week I had a few good predictions (Keenum, Taylor, and Dalton all had good weeks) and at least one bad one (I had Stafford at the bottom, but he ended up toward the top. No TDs though…). This week we’re still looking for streamers as it’s another byepocalypse week with 6 teams out.

Here are the models for this week:

ModelTypePlayerScoreOppScoreScoreOppScoreHomeAwayHomeAwayPlayerOppDataYears
ALMXXXXXX 2015-2017
BLMXXXXXXX2015-2017
CLMXXX 2015-2017
DGibbs SamplerXXX X2015-2017
EGibbs SamplerXXX XXX2015-2017
FPLSXXXXX X2015-2017
GPLSXXXXX X2017
HLMXXX X X2017
IMany LMsXXXXXX 2015-2017
JMany LMsXXXXXX 2016-2017*

 

Model A:

.Model A-1.png

Terms:

Player
Score
OppScore
Score*OppScore
Home/Away
Home/Away*Player (Some players have a bit more of a problem playing at home or away. Looking at you, Roethlisberger…)
Data years: 2015 - 2017

Model B:

.Model B-1.png

Terms:

Player
Score
OppScore
Score*OppScore
Home/Away
Home/Away*Player
Opponent
Data years: 2015 - 2017

Model C:

Terms:

Player
Score
OppScore
Data years: 2015 - 2017

Model D:

.Model D-1.png

Terms:

Player
Score
OppScore
Opponent
Data years: 2015 - 2017

Bayesian (Gibbs sampler)

Model E:

.Model E-1.png

Terms:

Player
Score
OppScore
Score*OppScore
Home/Away
Home/Away*Player
Opponent
Data years: 2015 - 2017

Bayesian (Gibbs sampler)

Model F:

.Model G-1.png

Terms:

Player
Score
OppScore
Score*OppScore
Score^2
OppScore^2
Home/Away
day (Thu/Sun/Mon)
Data years: 2015 - 2017

Model details: Partial least squares fit

Model G:

.Model G-1.png

Terms:

Player
Score
OppScore
Score*OppScore
Home/Away
Opp
Data years: 2017 only

Model details: Partial least squares fit

Model H:

.Model H-1.png

Terms:

Player
Score
OppScore
Home/Away
Opp
Data years: 2017 only

Model details: Linear model

Model I:

.Model I-1.png

Terms may include:

Player
Score
OppScore
Score*OppScore
Home/Away
Home/Away*Player
Opponent
Data years: 2015 - 2017

Model details: This model tries to predict all of the components of scoring separately (pass yards, pass TD, rush yards, 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

Model J:

.Model J-1.png

Terms may include:

Player
Score
OppScore
Score*OppScore
Home/Away
Home/Away*Player
Opponent
Data years: 2016-2017 for most people, 2017 only for Alex Smith and Jared Goff

Model details: This model tries to predict all of the components of scoring separately (pass yards, pass TD, rush yards, 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. I also modified the data years; 2016 and 2017 for most players, 2017 only for Alex Smith and Jared Goff since there’s been so much improvement since last year.

 

Weekly model summary
 

RankABCDEFGHIJ
1Deshaun WatsonDak PrescottDeshaun WatsonDrew BreesDrew BreesDeshaun WatsonDeshaun WatsonDrew BreesDeshaun WatsonDeshaun Watson
2Dak PrescottDrew BreesDrew BreesRussell WilsonDak PrescottRussell WilsonDrew BreesDak PrescottDrew BreesDak Prescott
3Drew BreesDeshaun WatsonRussell WilsonJameis WinstonRussell WilsonDrew BreesDak PrescottDeshaun WatsonDak PrescottAlex Smith
4Russell WilsonRussell WilsonDak PrescottDeshaun WatsonJameis WinstonCam NewtonRussell WilsonRussell WilsonRussell WilsonDrew Brees
5Cam NewtonJameis WinstonCam NewtonDak PrescottDeshaun WatsonDak PrescottAlex SmithCam NewtonJameis WinstonDerek Carr
6Jameis WinstonCam NewtonJameis WinstonCam NewtonCam NewtonBlake BortlesCarson WentzDerek CarrMarcus MariotaMarcus Mariota
7Blake BortlesBlake BortlesKirk CousinsMatthew StaffordBlake BortlesJameis WinstonMarcus MariotaKirk CousinsCam NewtonEli Manning
8Carson WentzDerek CarrDerek CarrDerek CarrDerek CarrKirk CousinsDerek CarrCarson WentzDerek CarrBlake Bortles
9Derek CarrMarcus MariotaMarcus MariotaBlake BortlesMatthew StaffordMarcus MariotaTyrod TaylorJared GoffBlake BortlesRussell Wilson
10Kirk CousinsKirk CousinsMatthew StaffordAlex SmithMarcus MariotaCarson WentzCam NewtonJameis WinstonMatthew StaffordKirk Cousins
11Marcus MariotaMatthew StaffordBlake BortlesKirk CousinsKirk CousinsDerek CarrJared GoffAlex SmithAlex SmithCam Newton
12Tyrod TaylorTyrod TaylorAlex SmithMarcus MariotaTyrod TaylorTyrod TaylorJameis WinstonDrew StantonKirk CousinsJameis Winston
13Andy DaltonC.J. BeathardTyrod TaylorTyrod TaylorAlex SmithMatthew StaffordMatthew StaffordJacoby BrissettTyrod TaylorMatthew Stafford
14Matthew StaffordAlex SmithCarson WentzJoe FlaccoCarson WentzAlex SmithBlake BortlesJay CutlerEli ManningMatt Ryan
15Matt RyanCarson WentzEli ManningEli ManningEli ManningEli ManningMatt RyanJosh McCownC.J. BeathardCarson Wentz
16Eli ManningEli ManningMatt RyanJay CutlerJosh McCownMatt RyanJay CutlerTyrod TaylorMatt RyanC.J. Beathard
17Alex SmithMatt RyanAndy DaltonJosh McCownMatt RyanJosh McCownBrett HundleyC.J. BeathardJosh McCownJosh McCown
18C.J. BeathardAndy DaltonJoe FlaccoMatt RyanJay CutlerAndy DaltonDrew StantonMatt RyanCarson WentzTyrod Taylor
19Josh McCownJosh McCownJosh McCownCarson WentzJoe FlaccoJay CutlerJosh McCownTrevor SiemianAndy DaltonTrevor Siemian
20Jay CutlerJoe FlaccoJay CutlerAndy DaltonJared GoffJoe FlaccoKirk CousinsBlake BortlesJared GoffAndy Dalton
21Joe FlaccoJay CutlerTrevor SiemianJared GoffAndy DaltonTrevor SiemianEli ManningMarcus MariotaJoe FlaccoJoe Flacco
22Jared GoffJared GoffC.J. BeathardTrevor SiemianTrevor SiemianJared GoffJoe FlaccoMatthew StaffordJay CutlerJared Goff
23Trevor SiemianTrevor SiemianJacoby BrissettJacoby BrissettC.J. BeathardJacoby BrissettJacoby BrissettEli ManningTrevor SiemianJay Cutler
24Jacoby BrissettJacoby BrissettJared GoffC.J. BeathardJacoby BrissettC.J. BeathardC.J. BeathardJoe FlaccoJacoby BrissettJacoby Brissett
25Drew StantonDrew StantonBrett HundleyBrett HundleyBrett HundleyDrew StantonTrevor SiemianAndy DaltonDrew StantonDrew Stanton
26Brett HundleyBrett HundleyDrew StantonDrew StantonDrew StantonBrett HundleyAndy DaltonBrett HundleyBrett HundleyBrett Hundley

Conclusions

[NOTES: I think I have all the right QBs in this week, but Beathard might become Garroppolo before the week is through. ]

I finally dropped Stafford and he had an OK week. Not great, but OK. I still feel okay about this, especially because my streamer of the week, Case Keenum, had a great week for me. Full disclosure: I have Goff already, so I might be higher on him than is warranted.

Let’s look at the top streaming options this week from Model H:

  • Jared Goff (at NYG): The number one fantasy QB this week, Jared Goff (I told you I’d be a little high on him), is up against the Giants who have been less than dominant against the pass this year. It’s a road game, with the notorious west-coast-team-playing-an-east-coast-team-in-the-morning curse, but I’m not worried about that as I don’t believe it exists. Well, that’s not true, I think it’s about a 1 (real) point difference. Check out the Data Exploration section of this blog for more. Still, I like him, he’s been a solid play, and the nominal score of 22.75 points isn’t bad.
  • Matthew Stafford (at GB): I know I’ve been hard on him, but against a mediocre Green Bay defense he’s not a bad choice. If they’re ahead in the end, though, he may not have to throw as much and perform a patented Matt Stafford 4th quarter comeback.
  • Jay Cutler (v OAK): Oakland hasn’t dome much to stop the pass, they’re getting back Devante Parker, and they just got rid of Jay Ajayi. They’re also slated to be playing from behind. A risky play, but if you’re out of options my models like Cutler this week.
  • Josh McCown (at BUF): low expected score (around 20 points) against a tough defense. Normally I’d avoid him, but he’s been getting it done, fantasy-wise, and turning in good performances. You could do worse. For example, any of the following guys that Model H likes:

CAUTION:

  • Jacoby Brissett (at HOU)
  • Drew Stanton (at SF)
  • C.J. Beathard (at ARI): Model H likes all these guys and puts them in the middle of the pack for the week, but that doesn’t really pass the smell test. All are predicted to get less than 20 points in road games against good defenses. In fact, these are all defenses I’d be streaming this week if I were you. Model H really likes them, but my next most-accurate models (D, E, and G) all say stay away.

This looks like a pretty rough week for streamers. Hopefully you’ve got one of those top 8 QBs. I’m also hoping for a poor performance from Russell Wilson since I’m going up against him this week, but a poor performance from him means 15 points instead of 30. This might get ugly for me.

Week 9 DEF Predictions

Week 9 K Predictions