Best Ball Strategy: Wide Receiver Back Stop Theory
What up, data degenz.
Welcome to the second post of the series Going In Raw. In this post we will be taking a look at best ball draft strategy for the wide receiver (WR) position.
For each Going In Raw post:
I’ll explain what lead me to the post and the intent
I’ll explain the data model and assumptions
I’ll provide the data with short analysis
I’ll end with a takeaway
Lead and Intent
If you are not familiar with best ball, it can be described as a draft like your standard redraft league without the management throughout the season. Each week, the highest scoring players drafted at each position(s) automatically start, which eliminates all day-to-day management decisions in your typical redraft league. Basically, you draft and then let the cards fall where they may.
It is not uncommon for a habitual daily fantasy sports (DFS) and tournament player to tout selecting volatile players with high ceilings. Why? Because to win a tournament, you need the value of the highest scoring players at each position for the week of play.
Best ball is extremely new to the marketplace and it seems to be increasingly thought that the same strategy for weekly DFS and tournaments applies to all formats of best ball. That is, draft volatile players and be rewarded for their ‘boom’ high ceiling weeks. After all, how great is it that you are able to draft a team of volatile players and not have to make the decision of when to start them to reap maximum benefits? It’s a great concept and all, you know, drafting volatile players that boom for 30 points every week.
When I think about this idea, I think about other factors that need to be considered.
For each team on any operating platform, there are a fixed number of positions that will start. What if you draft all volatile payers and they each have their highest scoring weeks on the same week of the season? Well, then you lose the value of the one week that you were banking player X to contribute to the overall teams score.
What if you select all volatile players and they all happen to have their lowest scoring week of the season the same week?
As much as I want to have a portfolio of volatile players, there is a significant amount of uncertainty to when the boom and bust weeks will happen. If I could have exposure to volatile players, but also hedge the volatility risk with high(er) floor consistent production, that would be ideal. The theory is to have a ‘backstop’ of production to minimize the impact of the floor weeks that come attached to volatile high ceiling players.
To test this theory, I decided to complete best ball simulations for the WR position. In these simulations, I constructed WR lineups that could be drafted today based on approximate ADP. As the WRs selected were based on real construction lineups that can be drafted in 2020, the fantasy points are based on what the results would have been in 2019. The kicker: I used Larry Fitzgerland and/or Cole Beasley as players within each construction. Why? Because they are both 2020 late round picks who are low ceiling players. Would there have been value in having either on your best ball team in 2019? Let’s look at the simulations.
Data Model
All raw data for player statistics (i.e. receptions, receiving yards, receiving TDs) is from pro-football-reference.com and points were converted into PPR scoring
The data is based on 2019 fantasy production
The data is only specific to the WR position
Excluding Sammy Watkins who played 13 games, the data only includes WRs that played => 14 games
The simulations were completed for all 17 weeks of the season
Typical best ball drafts extend from weeks 1 to 16. For this study I viewed the addition of week 17 as a larger data set to analyze
Three starting wide receiver positions was the construction chosen
Starting scores for each week are highlighted in green in each data set
Data & Analysis
Simulation 1
Snapshot Insight: This is an eight WR construction. Four of the wide receivers finished in the top 24 in 2019. Robby Anderson's second best week of the season in week 12 was not able to be utilized. Both Fitzgerland (8 weeks) and Beasley (6 weeks) combined for 20% of the WR production. Fitzgerald contributed more points than Tyler Boyd who finished 2019 as WR18.
Click all images to enlarge
Simulation 1A
Snapshot Insight: This is the same construction as simulation 1 with Lockett removed to see where his six contributed weeks worth of production would fall. Boyd gained one week (6 total), Slayton gained one week (5 total), Fitzgerald gained one week (9 total), Beasley gained three weeks (9 total), Fitzgerald and Beasley’s combined contributed value increased from 20% to 27%.
Simulation 2
Snapshot Insight: This is a nine WR construction. Four of the wide receivers finished in the top 24 in 2019. Cooper’s 22.6 point performance week 1 wasn’t able to be utilized. Beasley accounted for 11% of the WR production (7 weeks), which was nearly as much as teammate Brown who finished as the WR20. Three of Brown and Beasley’s scores were utilized the same week (weeks 4, 7, 16).
Simulation 3
Snapshot Insight: This is a nine WR construction. Five of the wide receivers finished in the top 24 in 2019. Williams was only utilized one week. Beasley again accounted for 11% of the WR production (7 weeks).
Simulation 4
Snapshot Insight: This is an eight WR construction. Five of the wide receivers finished in the top 24 and one in the top 30. Boyd’s 21.1 point performance week 12 wasn’t able to be utilized. Fitzgerald accounted for 9% of the WR production (6 weeks), which was more than Brown who finished as the WR20.
Simulation 4A
Snapshot Insight: This is the same construction as simulation 4 with Boyd removed to see where his 5 contributed weeks worth of production would fall. McLaurin gained one week (6 total), Landry gained one week (7 total), Brown gained two weeks (6 total), Fitzgerald gained one week (7 total). Fitzgerald’s contributed value increased from 9% to 11%.
Simulation 5
Snapshot Insight: This is an eight WR construction and an extreme example as six wide receivers finished in the top 24. Gallups 24.3 point performance week 5 wasn’t able to be utilized. Both Fitzgerland (3 weeks) and Beasley (5 weeks) combined for 13% of the WR production.
Simulation 5A
Snapshot Insight: This is the same construction as simulation 5 with Gallup removed to see where his six contributed weeks worth of production would fall. Fitzgerald gained two weeks (5 total), Beasley gained one weeks (6 total), Godwin gained one week (9 total), Chark gained one week (6 total), Boyd gained one week (8 total). Fitzgerald and Beasley’s combined contributed value increased from 13% to 16%.
Simulation 6
Snapshot Insight: This is a nine WR construction. Three of the wide receivers finished in the top 12 and two others in the top 30. Both Fitzgerland (4 weeks) and Beasley (5 weeks) combined for 14% of the WR production.
Simulation 6A
Snapshot Insight: This is the same construction as simulation 5 with McLaurin removed to see where his six contributed weeks worth of production would fall. Fitzgerald gained two weeks (6 total), Beasley gained one week (6 total), Robinson gained two weeks (10 total), Hopkins gained one week (9 total), Boyd gained one week (8 total). Fitzgerald and Beasley’s combined contributed value increased from 14% to 18%.
Beasley & Fitzgerald Rolled Up Simulations Contributions
Snapshot Insight: There were 10 total simulations completed. Beasley was in 8 of the simulations. On average, Beasley was utilized 6 of the 17 weeks and contributed 101.2 points or 10% of the overall WR construction points. Fitzgerald was in 6 of the simulations. On average, Fitzgerald was utilized 7 of the 17 weeks and contributed 102.7 points or 10% of the overall WR construction points.
Key Takeaways
The purpose of these simulations and analysis was to understand if players that have ‘low ceilings’ can add value to best ball WR construction. Acting as the low ceiling test subjects, both Fitzgerald and Beasley’s scores were utilized nearly 7 of the 17 weeks or 41% of the time in all 10 simulations.
The simulations completed were extreme WR construction scenarios as Fitzgerald and Beasley were competing with 4 or more top 24 2019 WRs in nine of the ten simulations. It was very surprising to find that both Fitzgerald and Beasley scores were used as much as they were. Prior to the analysis, I would not have thought that Cole Beasley and Larry Fitzgerald would have been utilized in 6 of the 17 weeks (each) on a construction team with Julio Jones, DeAndre Hopkins, Allen Robinson, and Tyler Lockett (Simulation 6A), but they were.
Based off these simulations, which is a very small sample size, it may be inferred that players that do not have the highest ceilings, but produce consistently week-to-week, may add value in best ball drafts. This ‘backstop production theory’ will need to be researched and tested further, however, the theory is that consistent week-to-week producers add value because of the higher point floor they offer compared to volatile players with lower point floors.
But who are these WRs and how many should be targeted?
My theory is that these WRs primarily play out of the slot. WRs that are consistently catching 4 to 5 receptions a game running underneath routes. According to PFF.com, here are the the players that ran the greatest % of snaps out of the slot in 2019 (=>100 total snaps ran):
The strategy I am testing this year is to target drafting 8-10 WRs total on a 20 man roster. Of the 8 to 10 wide receivers, I am drafting one to two of these slot wide receivers while mixing in other volatile players. In my eyes, that is a nice 80% - 90% higher ceiling risk/10% - 20% hedge approach.
In no particular order, my favorite 10 players on the slot WR list include: Randall Cobb (now in HOU), Larry Fitzgerald, Golden Tate, Anthony Miller, Danny Amendola, Jamison Crowder, Tyler Lockett, Julian Edelman, Cooper Kupp, and COLE BEASLEY.
As of 7/26, I am sitting on 28% Cole Beasley exposure on Drafters.com. I will let you know how the theory plays out.
Note: I believe this theory applies mostly to <= 12 man leagues. In tournaments, I am only mixing in these late round players (i.e Amendola) in when stacking (i.e. Stafford/Amendola).
