Every week DraftKings pays out big prizes to winners of their contests. Most sites focus on what people did, and pick apart the best lineup. I’m not going to do much of that here. Instead I’ll show you what was the best lineup possible and analyze the strategies of top DFS players.
- BillyVonElds wins with 253.72 points
- Read his Twitter
- 20 Lineups Entered
- Predicted breakout games for certain players
- Won on a Drew Brees-Tre’Quan Smith stack but actually played more of Carson Wentz-Nelson Agholor and Cam Newton-D.J. Moore
- Defense DOES NOT Get Better as the Season Goes On!
- Some coaches say it and you may think it about the Saints but it’s not true
- T.Y. Hilton was the best pick at wide receiver
- All of our better than winner lineups used T.Y. Hilton
- 149.00 to cash
- Best possible lineup: 278.72 points and a salary of $49,900
BillyVonElds Wins with 253.72 Points
He won with a Brees-Tre’Quan Smith stack but he was more into Carson Wentz-Nelson Agholor and Cam Newton-D.J. Moore. Take a look at his player pool.
|QB||QB Percentage||RB||RB Percentage||WR||WR Percentage||TE||TE Percentage||DST||DST Percentage|
|QB Cam Newton||20||RB Saquon Barkley||84||WR Nelson Agholor||49||TE Jordan Reed||24||DST Chargers||39|
|QB Carson Wentz||20||RB Dion Lewis||39||WR D.J. Moore||44||TE Ricky Seals-Jones||24||DST Texans||34|
|QB Drew Brees||14||RB Ezekiel Elliott||34||WR Tre’Quan Smith||39||TE James O’Shaughnessy||20||DST Panthers||10|
|QB Lamar Jackson||14||RB Christian McCaffrey||20||WR Corey Davis||29||TE Luke Willson||10||DST Steelers||10|
|QB Eli Manning||14||RB Alvin Kamara||20||WR Odell Beckham Jr.||29||TE O.J. Howard||5||DST Redskins||5|
|QB Josh Rosen||10||RB David Johnson||14||WR Seth Roberts||29||TE Evan Engram||5|
|QB Ryan Fitzpatrick||5||RB James Conner||14||WR Alshon Jeffery||24||TE Greg Olsen||5|
|RB Melvin Gordon III||10||WR T.Y. Hilton||20||TE Jeff Heuerman||5|
|RB Alex Collins||10||WR Kenny Golladay||20|
|RB Mark Ingram||5||WR Michael Thomas||14|
|WR Mike Evans||10|
|WR Larry Fitzgerald||10|
|WR DeAndre Hopkins||10|
|WR Golden Tate||10|
Very hot on Saquon Barkley, way to take a big risk and get rewarded.
Also notice that the first 3 WR’s were all risky WR picks that were all stacks with his top 3 QB’s. BUT the player pool doesn’t tell the whole story.
Looking at his stacks, we see Brees-Smith were used in 10% of lineups and Brees-Thomas were used in 10% as well. He played a lineup with both of them in a triple stack, and then 1 with each of them by themselves.
Brees-Ingram were never used but Brees-Kamara was a stack. Probably because of the pass catching prowess of Alvin Kamara
Similar story with Wentz and Agholor, Jeffery, and Golden Tate.
Here we see Lamar Jackson used 3 times out of 20 but never with Alex Collins. This is probably deliberate.
Cam Newton Stacks
He stacked Cam with D.J. Moore in 3/4 Cam Lineups. And Cam with McCaffrey in half of those. BillyVonElds definitely views Ingram and Collins as different style running backs than McCaffrey.
From how complicated the overlap and stacks/anti-stacks are I’d reckon this guy used some manual intervention to enter these lineups.
On the season BillyVonElds has spent up to $400 each week.
Nice win. Good luck to BillyVonElds in the future.
My Thoughts on GuruElite
I’m a bit skeptical of the “Elite Mafia” and Tommy G, Jeff Mans etc.
— Billy Elder (@BillyVonElds) November 19, 2018
My opinion is that GuruElite’s success is inflated by the amount of followers they have, nearly 40,000! If Tommy G or GuruElite was so great he should be winning more, or at least paying more than $100 a week to try.
Take a look at TommyG’s performance.
TommyG1979 2018 Milly Maker Performance
This is worse than VonElds in just about every week. And TommyG certainly doesn’t throw down in the Milly Maker. He seems more like the guy that covers the board in roulette and brags about the chips he gets back (it looks like he bets on a ton of different sports based on his twitter). Still, I haven’t analyzed enough data on TommyG to be sure either way. But based on what I’ve seen this far I think he’s significantly overhyped.
Optimal Lineup: 278.72 Points
A bit of a lower scoring week. The top possible lineup on the main slate was
|New Orleans Saints||DST||13.00||2100.0||NO|
The Saints D was a big shocker this week. Many thought Carson Wentz would do well against a D that has let up a lot of big scores this year. But keep in mind mean reversion. The Saints were projected to have a good D in the offseason and have been under performing.
Defense DOES NOT Get Better as the Season Goes On!
Some people say that defenses get better as the season goes on. And while that could be true it doesn’t hold for points allowed or for DraftKings points scored.
Below I’ve plotted the median in blue and average in orange.
In the case of the Saints D injuries and mean reversion were the most likely the biggest factors.
Better Than Winner Player Pool
There were at least 2,308 lineups better than the winner’s 253.72. The player pool of these better than winner lineups is shown below
|New Orleans Saints||0.282||NO||DST||PHI||2100.0||13.00|
Thanks to Jason of The Perfect Lineup Podcast for the suggestion to add salaries to the above table.
T.Y. Hilton was the unanimous best player in the better than winner player pool (we use a pruning algorithm to ensure the player pool contains the optimal lineup. This pruning does cut down on the number of lineups we generate but guarantees we get the best score).
Drew Brees was the best QB and blast from the bast, Antonio Gates and Jordan Reed we’re the best TE’s.
|Player||Number of Lineups||Paid||Won||win/entry|
Mallen21, a suspected tout or overrated player by one of our readers, has another good showing. He’s on the list of players we’ll review further but so far he is looking good.
Slaytations, the Week 5 winner and a reader of this blog, did well again. Here’s a look at how all the 2018 Milly Maker Winner’s fared in Week 11.
|Week Winner||Player||Number of Lineups||Paid||Won|
jab76r cashes again. Nickelback4Lyfe (another reader) continues to max enter. He’s come close to cashing most weeks and seems to outperform “pro’s” like Awesemo so keep at it and good luck Nickelback4Lyfe.
From what I could gleam from GuruElite last week, D.J. Moore was one of their big picks. Tommy G played D.J. Moore in 2 lineups but was actually more into Tyrell Williams, playing Williams in 3 lineups and 2 lineups for Moore. Tommy G only entered 5 lineups though so it’s hard to gleam much. He lost money for the 10th time this year on the Milly. Maybe he’ll have a big hit soon but 2018 is a far cry from 2017 when he would enter closer to 100 if not 150 lineups.
Pro vs Herd
Here is the section where I compare the Rotogrinder’s Top 50 with everyone else (their rankings are heavily skewed to high volume players but it’s the best we have… for now).
chess_is_ok, awesemo, and chipotleaddict all are starting to look like permanent fixtures on the bottom of these standings.
cubasugar1 takes home the worst lineups prize. Created 150 lineups and lost on all of them 🙁 Go pour yourself a legitimate Havana Club instead of that Bacardi version and I bet you’ll feel alright.
Pro vs Herd Point Distribution
The rg_top_50 did better on average but it was the other guys, the “herd” that took home top prize.
Pro vs Herd Player Pool
|lineup_hash||Percentage Pro||Percentage Herd||pro_favoritism|
|Melvin Gordon III||6||7||-1.0|
|Odell Beckham Jr.||23||19||4.0|
In spite of Corey Davis and James Conner the average pro score beat the herd. Barkley and Saints probably played a big factor in that. The Saints we’re a really good pick.
On the other hand, the herd was hot on the Cowboys. Elliott was a good pick but Cooper was not. Same goes for the Cardinals (I had the Texans and Cardinals D last week so I definitely got burned a bit too).
T.Y. Hilton was the biggest pick by the herd.
Happy Thanksgiving. When you miss a player in your lineups be thankful you’re not a turkey.
The above meme is not true but I hope it made you laugh.
Speaking of Turkeys A Note on Inference
People like to think cause and effect is something you can tell with certainty. Spoiler alert, you can’t.
Consider the weight of a turkey. Starts off around 5lbs and keeps growing each year until… BAM dead turkey. It’s Thanksgiving and it goes to 0 as people devour it.
How the hell do you predict that?
Sure you might understand turkey and thanksgiving but many data problems are the equivalent of forecasting the turkey’s weight when all you have is a chart of turkey weight (looking at you financial markets). Be wary about inferences you take and trust you place in forecasting models like predictions.
This example was inspired by one of Nassim Taleb’s books who I believe had his example inspired by someone else.