DraftKings Week 11 Recap

RobKaufman's Winning Lineup

Week 11 is in the past. This week’s $1MM winner was RobKaufman. He entered 50 lineups and won with a millionaire maker outlier of 221.18 points.

RobKaufman's Point Distribution

Only 6 of his lineups cashed! (145.36 to cash in Week 11)

What a week for this guy. Here’s the usual look at how he engineered his lineups.

RobKaufman’s Winning Strategy

Player pool 49 players (this is somewhat similar to Al_Smizzle’s approach and suggests that Rob manually highlighted a subset of players).

His most used players were

Percentage Player
63 Chris Thompson
63 Rob Gronkowski
61 Chiefs
57 Jeremy Maclin
57 Jay Cutler
47 Jarvis Landry
45 Kareem Hunt
39 Jamison Crowder
35 Cooper Kupp
27 Jordan Howard
25 Robert Woods
21 Brandin Cooks
17 Michael Thomas
17 Jared Goff
15 Jaguars
15 Chargers
15 Bruce Ellington

Note that even Jay Cutler’s dreadful Week 11 performance (3 picks at half) could sink RobKaufman.

Jay Cutler Don't Care

His largest stack was a RB-DST stack at 39% Kareem Hunt-Chiefs. Next up was Jay Cutler-Jarvis Landry at 33%.

Percentage Player
41 Jeremy Maclin, Rob Gronkowski
39 Chiefs , Rob Gronkowski
39 Chiefs , Kareem Hunt
37 Jay Cutler, Rob Gronkowski
37 Chiefs , Chris Thompson
37 Chris Thompson, Jarvis Landry
35 Chris Thompson, Jay Cutler
33 Chris Thompson, Jeremy Maclin
33 Jarvis Landry, Jay Cutler
33 Chris Thompson, Rob Gronkowski
31 Chiefs , Jeremy Maclin
31 Kareem Hunt, Rob Gronkowski
31 Cooper Kupp, Rob Gronkowski
29 Jay Cutler, Jeremy Maclin
29 Chiefs , Jay Cutler
27 Jeremy Maclin, Kareem Hunt
27 Jarvis Landry, Rob Gronkowski

Now it’s possible those two stacks were deliberate but it’s tough to tell given the consequential frequent pairs. The biggest takeaway I see is that RobKaufman won by using a engineered player pool rather than adhering to common lineup stacks.

Congratulations RobKaufman.

Business Category (English) 468x60

A Look at DFS Pros

As a group, the Rotogrinders top 50 entered 4,399 lineups accounting for 2.4% of all entries in Week 11’s Millionaire Maker. “TwoGun” was the highest placing “pro” with 198.66. ChipotleAddict finished 8th among pro’s with 191.68.

The top 50 lost $41,670 with only 3 pro’s making money on the contest.

payout payout per lineups entered
screen_name
ianj300 170.0 12.142857
aejones 95.0 4.523810
twogun 90.0 2.727273
0_for_5 -10.0 -5.000000
iampgm -40.0 -20.000000
grinderscal -50.0 -2.500000
shocae -90.0 -15.000000
dinkpiece -100.0 -5.000000
eriii -120.0 -20.000000
bigt44 -120.0 -20.000000
csuram88 -210.0 -17.500000
araven52 -230.0 -17.692308
bales -320.0 -16.000000
al_smizzle -330.0 -2.200000
bidle -340.0 -2.266667
hen-dog -365.0 -9.125000
tommyg1979 -435.0 -8.055556
bric75 -660.0 -4.400000
boggslite -670.0 -9.571429
kingfi -740.0 -4.933333
anilprao88 -865.0 -5.766667
moklovin -920.0 -6.133333
str8_cash_homies -1015.0 -6.766667
brandonadams -1060.0 -7.066667
chipotleaddict -1125.0 -7.500000
ehafner -1200.0 -8.000000
damp -1210.0 -9.307692
spx333 -1230.0 -13.369565
flavorflav -1240.0 -8.266667
cubsfan333 -1260.0 -14.318182
chess_is_ok -1365.0 -9.100000
youdacao -1375.0 -9.166667
nolesman -1385.0 -9.233333
scout326 -1395.0 -9.300000
papagates -1445.0 -9.633333
wakeywakey -1515.0 -10.100000
awesemo -1650.0 -11.000000
rikkidee -2015.0 -13.433333
bdholla89 -2050.0 -13.666667
shawnzhan -2075.0 -13.833333
tomjk321 -2215.0 -14.766667
petrgibbons -2455.0 -16.366667
fallfromgrace -2500.0 -16.666667
thatstunna -2630.0 -17.533333

Note: Grinderscal and cubsfan333 won earlier this year and are still playing. Also, this chart is ranked by winnings. I’d argue Bales and CSURAM88 did worse than Al_Smizzle because they entered fewer lineups and on average lost way more per lineup.

Summary statistics for the pro’s

Points
count 4399.000000
mean 131.135826
std 20.266220
min 61.400000
25% 117.330000
50% 130.560000
75% 145.400000
max 198.660000

The most common pro stacks used Tom Brady

Percentage Player
18 Jamison Crowder, Kareem Hunt
13 Jamison Crowder, Michael Thomas
11 Brandin Cooks, Tom Brady
11 Rob Gronkowski, Tom Brady
11 Kareem Hunt, Michael Thomas
11 Drew Brees, Michael Thomas
10 Kareem Hunt, Rex Burkhead
10 Bruce Ellington, Kareem Hunt
10 Chris Thompson, Michael Thomas
10 Chris Thompson, Kareem Hunt
10 Jaguars , Kareem Hunt
10 DeAndre Hopkins, Kareem Hunt
9 Brandin Cooks, Kareem Hunt
9 Kareem Hunt, Tom Brady
9 Chris Thompson, Jamison Crowder
9 Jaguars , Jamison Crowder
9 Jamison Crowder, Rob Gronkowski

Pro’s We’re Tracking

Bales

Jonathan Bales entered 20 lineups again. He lost $320 and did worse than the average non pro player.

Points
count 20.000000
mean 126.689999
std 17.515501
min 89.319990
25% 114.300000
50% 126.220000
75% 138.410000
max 164.460000

He fell into the Rex Burkhead trap, using him half of his lineups

Percentage Player
49 Rex Burkhead
44 Chris Thompson
39 Jamison Crowder
34 Bruce Ellington
29 Amari Cooper
29 Michael Thomas
29 Melvin Gordon
29 Michael Crabtree
24 Lions
24 Vernon Davis
24 Kareem Hunt
24 A.J. Green
19 LeSean McCoy
19 Tom Brady
19 Brandin Cooks
19 Rob Gronkowski
19 DeAndre Hopkins

But the Lion’s Defense was a solid pick. He went super heavy on Tom Brady-Rob Gronkowski and was really hoping they’d go off.

Percentage Player
29 Jamison Crowder, Michael Thomas
19 Jamison Crowder, Rex Burkhead
19 Amari Cooper, Rex Burkhead
19 Amari Cooper, Lions
19 Lions , Rex Burkhead
19 Amari Cooper, Brandin Cooks
19 Rob Gronkowski, Tom Brady
19 Michael Crabtree, Rex Burkhead
19 A.J. Green, Chris Thompson
14 Bruce Ellington, Jamison Crowder
14 Bruce Ellington, Chris Thompson
14 Amari Cooper, Chris Thompson
14 Bruce Ellington, Tom Brady
14 Amari Cooper, Tom Brady
14 Lions , Tom Brady
14 Rex Burkhead, Tom Brady
14 Brandin Cooks, Lions

Unfortunately, Tom Brady-Brandin Cooks was the better play. NOTE: this is why checking pairs of players is important. Both Brandin Cooks and Rob Gronkowski were used the same amount but because Gronk and Brady were always played together, Bales missed out on Brandin Cooks.

This begs the question: Why doesn’t Bales play more lineups? He’s written books about DFS and sold a company (Fantasy Labs) but won’t play more than 20 lineups each week in the $1MM Maker. If he added 4 lineups and played Cooks in all of them, to mirror his Gronk utilization, he would have been able to capitalize on the predictable event that some Patriot’s player would catch a TD and rack up yards.

CSURAM88 Only Cashed 1 Lineup out of 12

Glossing over him this week. He played like Bales with a Gronk-Burhead heavy player pool.

Tommyg1979 only enters 54 Lineups

He lost $435 and played less than the week before. Maybe he has some Christmas shopping to do or is focusing on other contests.

Business Category (English) 468x60

TommyG outplayed the herd but didn’t beat the rake

Points
count 54.000000
mean 137.831481
std 15.935921
min 101.360000
25% 127.085000
50% 138.310000
75% 149.340000
max 174.380000

He had Tyreek Hill in 95% of his lineups.

Percentage Player
95 Tyreek Hill
68 Rob Gronkowski
55 Alvin Kamara
53 LeSean McCoy
36 Tom Brady
31 Brandin Cooks
29 Bruce Ellington
21 Tyler Kroft
18 Rex Burkhead
14 Jaguars
12 Blaine Gabbert
12 Larry Fitzgerald
12 Austin Ekeler
12 Mike Wallace
12 Jared Cook
12 Michael Crabtree
12 Amari Cooper

Gronk didn’t help either. TommyG had Brady-Gronk in 36% of his stacks.

ChipotleAddict Lost $1,125 on 150 Lineups

Like TommyG, ChipotleAddict beat the herd but not the rake.

Points
count 150.000000
mean 131.861867
std 20.972767
min 80.900000
25% 117.165000
50% 131.230000
75% 146.955000
max 191.680000

He had a nice little bump around 175 points but it wasn’t high enough this week.ChipotleAddict Week 11 Point Distribution

LeSean McCoy was his top RB.

Percentage Player
53 DeAndre Hopkins
46 Michael Thomas
36 Jamison Crowder
33 LeSean McCoy
28 Vernon Davis
26 Emmanuel Sanders
25 Jordan Howard
25 Jerick McKinnon
24 Chris Thompson
24 Joe Mixon
23 Kirk Cousins
22 Todd Gurley II
19 Jaguars
17 Demaryius Thomas
16 Kareem Hunt
16 Rob Gronkowski
16 Golden Tate

And ChipotleAddict was hot on the Redskins offense. His highest stacks used Kirk Cousins (other pro’s favored Tom Brady)

Percentage Player
21 Jamison Crowder, Michael Thomas
20 Kirk Cousins, Michael Thomas
18 DeAndre Hopkins, Michael Thomas
17 Michael Thomas, Vernon Davis
16 DeAndre Hopkins, LeSean McCoy
16 DeAndre Hopkins, Jordan Howard
14 LeSean McCoy, Michael Thomas
14 Jamison Crowder, Kirk Cousins
13 Jamison Crowder, LeSean McCoy
13 Kirk Cousins, Vernon Davis
13 DeAndre Hopkins, Vernon Davis
13 DeAndre Hopkins, Emmanuel Sanders
13 DeAndre Hopkins, Joe Mixon
12 DeAndre Hopkins, Jamison Crowder
12 Chris Thompson, DeAndre Hopkins
12 DeAndre Hopkins, Jerick McKinnon
12 Jerick McKinnon, Michael Thomas

So good work picking Kirk Cousins over Tom Brady but it wasn’t enough for ChipotleAddict.

Al_Smizzle: Mr. Player Pool Continues His Strategy

He lost $330 on $3,000, putting him among the best as a high volume player. Al_Smizzle used Rex Burkhead the most but was saved by Brandin Cooks in 50% of his lineups.

Percentage Player
54 Rex Burkhead
50 Brandin Cooks
43 Jamison Crowder
40 Travis Kelce
36 Kareem Hunt
34 Jaguars
32 Michael Crabtree
32 Jeremy Maclin
30 Bruce Ellington
30 Benjamin Watson
28 Chris Thompson
27 Melvin Gordon
26 Jarvis Landry
26 Mark Ingram
25 Tom Brady
25 Rob Gronkowski
25 Evan Engram

He only used 41 different players. As someone who uses optimizers, this makes me believe he puts less faith in optimizing lineups versus generating lineups subject to constraints.

For those of you who aren’t that familiar with optimizers, typically you’d take the top 150 lineups ranked by predicted score or, better yet, predicted lineup “portfolio” performance. By reducing his player pool to 41 players, there are still about 200,000 valid lineups in a given week. But if you’d use 95 players there are 10+MM possible lineups. While 200,000 is a big number, it allows Al_Smizzle to not blindly trust an optimizer that might underplay a high upside player like Brandin Cooks because Rob Gronkowski is projected 1 higher point. The flipside is that picking the best player pool is tough.

Al_Smizzle’s NFL strategy is a lot of work to replicate if you play every day like in the NBA but may be the best approach for NFL where predictions are harder.

Here’s his stacks

Percentage Player
31 Brandin Cooks, Rex Burkhead
24 Jamison Crowder, Travis Kelce
22 Jamison Crowder, Rex Burkhead
20 Brandin Cooks, Jamison Crowder
20 Rex Burkhead, Travis Kelce
20 Brandin Cooks, Jaguars
19 Jaguars , Rex Burkhead
18 Brandin Cooks, Travis Kelce
18 Brandin Cooks, Tom Brady
18 Kareem Hunt, Rex Burkhead
17 Jaguars , Jamison Crowder
17 Brandin Cooks, Kareem Hunt
17 Evan Engram, Rex Burkhead
16 Michael Crabtree, Rex Burkhead
16 Bruce Ellington, Travis Kelce
16 Jamison Crowder, Melvin Gordon
16 Jarvis Landry, Kareem Hunt

Nothing jumps out as overly deliberate, which continues to suggest that Al_Smizzle spends the majority of his time finding ~40 players to use in all of his lineups versus optimizer/lineup engineering.

I’d guess Al_Smizzle knows the most about football out of all the DFS Pro’s.

Programming Category (English)468x60

Notable No Shows

CONDIA missed the last few weeks. Why?

On the other hand, SaahilSud might be back again after entering 1 lineup and cashing with 153.36. He used a hefty dose of the Patriots.

RB Alvin Kamara RB Kareem Hunt WR Roger Lewis DST Chargers WR Amari Cooper TE Jared Cook QB Tom Brady FLEX Rex Burkhead WR Brandin Cooks

The Herd Meta

The herd won the $1MM maker in Week 11. But on average, performed worse.

index Points_herd Points_pro
count 178542.000000 4399.000000
mean 128.658417 131.135826
std 21.688318 20.266220
min 0.000000 61.400000
25% 113.860000 117.330000
50% 128.520000 130.560000
75% 143.360000 145.400000
max 221.180000 198.660000

Here’s the herds player pool compared with Rotogrinder’s Top 50

lineup_hash Percentage Pro Percentage Herd pro_favoritism
Travis Kelce 12 22 -10
Melvin Gordon 12 19 -7
Jared Cook 6 13 -7
Tom Brady 18 25 -7
Derek Carr 10 15 -5
Larry Fitzgerald 7 12 -5
Brandin Cooks 22 27 -5
Tyreek Hill 6 11 -5
Michael Crabtree 19 24 -5
Jaguars 20 24 -4
Jeremy Maclin 14 17 -3
Alex Smith 7 10 -3
Alvin Kamara 16 19 -3
Jarvis Landry 10 13 -3
Jerick McKinnon 5 8 -3
Rams 5 8 -3
Amari Cooper 17 20 -3
Emmanuel Sanders 6 7 -1
Stefon Diggs 5 6 -1
Rex Burkhead 23 23 0
Chargers 14 14 0
Rob Gronkowski 25 25 0
Ted Ginn Jr. 5 5 0
Mark Ingram 11 10 1
Golden Tate 7 6 1
Bruce Ellington 19 17 2
Browns 7 5 2
Keenan Allen 7 5 2
Ravens 8 6 2
Mike Evans 11 9 2
Tyler Kroft 13 10 3
Chris Thompson 28 25 3
Lamar Miller 9 6 3
Evan Engram 17 13 4
DeAndre Hopkins 19 15 4
Jordan Howard 12 7 5
Drew Brees 14 9 5
Kirk Cousins 13 7 6
Vernon Davis 16 9 7
Todd Gurley II 14 6 8
Kareem Hunt 43 35 8
LeSean McCoy 20 10 10
Jamison Crowder 35 21 14
Michael Thomas 29 13 16
Joe Mixon 12 <5% <5%

The pro’s did better identifying Lesean McCoy and Kirk Cousins but missed on Michael Thomas and Joe Mixon.

Like the pro’s, the herd used Tom Brady in the majority of their stacks.

Percentage Player
12 Brandin Cooks, Tom Brady
12 Rob Gronkowski, Tom Brady
9 Brandin Cooks, Kareem Hunt
9 Rex Burkhead, Tom Brady
9 Jamison Crowder, Kareem Hunt
8 Brandin Cooks, Chris Thompson
8 Jaguars , Kareem Hunt
8 Kareem Hunt, Tom Brady
8 Brandin Cooks, Rex Burkhead
8 Kareem Hunt, Rex Burkhead
8 Bruce Ellington, Kareem Hunt
8 Kareem Hunt, Michael Crabtree
7 Chris Thompson, Rob Gronkowski
7 Chris Thompson, Kareem Hunt
7 Jeremy Maclin, Kareem Hunt
7 Rex Burkhead, Rob Gronkowski
7 Brandin Cooks, Rob Gronkowski

The best and worst non-pros.

payout payout/num_entries
screen_name
robkaufman 999190.0 19983.800000
koolrr 99980.0 99980.000000
moneymer 49980.0 49980.000000
gphoodrat 29920.0 7480.000000
nawmsayin 21560.0 143.733333
dagodfather -2710.0 -18.066667
nicky6 -2720.0 -18.133333
cuseguy51 -2800.0 -18.666667
jakz101 -2880.0 -19.200000
isupol -2970.0 -19.800000

isupol took the worst spot losing $2,970 on 150 lineups. Ouch.

Best Possible Lineup: 237.08 Points

This week, there were at least 1,645 better lineups than RobKaufman. The player pool of these better lineups with utilization percentages is below

Name Percentage
Kenny Stills 1.000
Brandin Cooks 0.895
Ricky Seals-Jones 0.798
Kirk Cousins 0.695
Davante Adams 0.536
LeSean McCoy 0.498
Baltimore Ravens 0.489
Los Angeles Chargers 0.459
Adam Thielen 0.452
Samaje Perine 0.421
D’Onta Foreman 0.419
Latavius Murray 0.384
Larry Fitzgerald 0.257
Blaine Gabbert 0.193
Mike Wallace 0.140
Dion Lewis 0.130
Ryan Grant 0.126
O.J. Howard 0.099
Adam Shaheen 0.095
Alex Collins 0.095
Mark Ingram 0.086
Tarik Cohen 0.078
Corey Coleman 0.077
Jordan Howard 0.075
Alvin Kamara 0.075
Bruce Ellington 0.074
Andy Dalton 0.063
Alex Erickson 0.061
Devontae Booker 0.052
Tampa Bay Buccaneers 0.047
Zay Jones 0.029
Travaris Cadet 0.025
Tom Savage 0.025
Brock Osweiler 0.024
Mike Williams 0.012
Travis Kelce 0.009
New York Giants 0.005
Rex Burkhead 0.002

Kenny Stills was the only unanimous perfect pick and Kirk Cousins was the best QB.

The best lineup was

player_name position score salary team_name
Kirk Cousins QB 28.98 6100.0 WAS
LeSean McCoy RB 28.60 7800.0 BUF
D’Onta Foreman RB 23.00 4000.0 HOU
Brandin Cooks WR 29.90 6600.0 NE
Davante Adams WR 23.60 5300.0 GB
Kenny Stills WR 34.00 4000.0 MIA
Ricky Seals-Jones TE 20.40 2500.0 ARI
Samaje Perine FLEX 22.60 3800.0 WAS
Baltimore Ravens DST 26.00 3400.0 BAL
 Totals 237.08 43500

Announcements

After the NFL regular season is over we will be working on a few articles. So far we’re planning

  • A look at how $1MM players do after winning the big prize
  • How much to cash every week
  • Pro and Herd Short Term Trends (how things like recent performance affecting the two groups differently)
  • Our thoughts on DraftKings GPP’s being skill or luck
    • Past pro’s like Bales, CSURAM88, CONDIA, and SaahilSud don’t appear to do significantly better than the other $1MM players
  • NBA Recaps

I know last time I said we’d finish the WR post for How Sharks Approach DFS but it’s going to take another week #ThanksgivingCreptUp

Good luck in Week 12. Our fade of the week is Todd Gurley II.

Our past recaps can be found archived here. We also have a “How Sharks Approach DFS Series” and if you really love these please consider supporting us on Patreon and sharing with your friends using one of the buttons below. 

Matteo Hoch

Matteo Hoch is the founder of Sports Data Direct and maintains a personal blog called Ergo Sum where he writes about data visualization and coding. You can generally find him in the DraftKings Sunday $3 early only contests under username mhoch2

2 Responses

  1. Adrian P says:

    Thanks for doing these. I’m kind of new at trying out Draftkings and I’m very interested in the types of strategies everyone else uses. Great work.

Leave a Reply

Your email address will not be published. Required fields are marked *