DFS Strategy: Roster Construction & How to Use Roster Percentages
Continuing our month of getting you ready for the 2021 DFS season, the most recent episode of the Fantasy Footballers DFS Podcast poured more gasoline on the fire of our DFS Strategy rollout. In DFS Roster Construction + How to Use Roster Percentages, Matthew Betz and myself walked through the roster construction process of DFS and why gaining this understanding matters so much to being successful long-term.
What Do We Mean By Roster Construction?
Roster Construction isn’t just about “playing the best plays”. If DFS were that simple, we wouldn’t need to strategize each week. The best DFS players understand the nuances of roster construction and know there is more to this than matchups and finding value in salaries.
When we speak of roster construction, we’re asking “how do these pieces fit together?”
You can easily find the ‘multiplier’ of a player by taking his projected fantasy points and dividing by his salary (i.e. 25 points/$5,000 salary = 5x multiplier). But there is another variable at play in this equation known as Roster Percentages. As I recently mentioned, DFS introduces a different level of strategy:
DFS is a different animal…
Josh Allen was the No. 1 QB in fantasy last year.
— Kyle Borgognoni (@kyle_borg) July 19, 2021
Maybe I cherry-picked that statistic but the point is we can’t simply apply conventional redraft, dynasty, or Bestball strategy when how your roster is put together and who else has your roster has to be part of the equation.
How can you be better than the field in terms of roster construction?
We need to initially think about positions differently. I covered this in full in How to Approach Each Position in DFS. The thrust of this article is going to focus on these action steps:
- Create a Roster Construction Checklist
- Create Player Pools
- Learn how to leverage Roster Percentages
Let’s begin with the ‘checklist’ and hopefully build from there to give you a better perspective of roster construction.
Roster Construction Checklist
There are a number of different “templates” for creating a lineup. This is not a perfect exercise but as you hand build lineups for DFS, keeping these high-yielding templates in mind and after you select your players asking yourself if you are hitting a lot of these criteria. But most of the winning lineups/experienced players in DFS start with this in mind:
- Do you have a Team Stack?
- Do you have a Game Stack?
- Is your QB paired with a pass catcher?
- Is your RB paired with their DST?
- Do you have only one RB from the same game?
- Did you pay up or punt at TE? And is your TE paired with your QB?
- Do you have at least one player who is “low-rostered”?
- Are any of your players so popular that you can’t create meaningful leverage?
I looked at every Milly Maker winning lineup in 2020 and the data is fairly obvious:
# of Winning Milly Maker Lineups
QB + Pass Catcher
This week on the Fantasy Footballers DFS Podcast we are spending a whole episode on Stacking so a follow-up article is in order. Beyond the stacking, some of the other “templates” show up around 50-60% of those winning lineups. The one RB from each game does have a high hit rate that we recommend factoring in your roster construction each main slate. We don’t recommend employing the RB/DST combo very often as its correlation isn’t nearly as strong (0.1) as some people think.
Creating Player Pools
This is one of the best parts of your DFS week and something I begin usually on Wednesdays. Call me old school but I personally love using a pen and paper to simply write down the matchups to get a bird’s eye view of the week.
Work through the slate game-by-game identifying players you like and dislike. Don’t even worry about salaries or projections yet. This is about identifying your priors and getting them “out of your head” and onto a sheet of paper. As noted in the Forecasting 101 article, priors don’t only have to be viewed as a negative thing. Instead, consider this a warm-up and low-level entry point to the week.
From that list, I begin to connect these players with salaries to find inherent values or those that seem like cash game players for the week. When finding
- QBs– 3 or 4
- RBs– 4 or 5
- WRs– 6 or more
- TEs– 2 or 3
- DEFs– 3 or 4
This is the beginning of your “player pool” but in order to start to find optimal lineups based on salaries and projections, we also must take this a step further and compare our priors with projected roster percentages found in the DFS Pass.
First, let’s go over a couple of simple rules of thumb for roster percentages. First, this is often called ‘ownership percentage’ on other sites but we prefer to make a slight adjustment in our language to roster percentage. These are guides based on algorithms but not perfect forecasting. In our DFS Pass, we will showcase a brand-new feature for 2021 called the Roster Percentage Report.
People are also emotional and the week has an ebb-and-flow to it. Players such as Alvin Kamara and other RBs can often show up on 70+ percent of rosters as situations change late and the best plays become the chalk plays. We need to remember that ‘safety’ is a flawed way of looking at DFS in tournaments. Every player has a downside.
Here are a couple of common terms used in regard to roster percentage projections:
- The Chalk– these are the most popular plays of the week and the conventional high floor/high ceiling plays. Chalk isn’t bad but when someone starts to creep above 15 percent and even 30 percent, you have to make a call on whether it’s a good week to ‘eat the chalk’.
- Low-rostered– We consider someone ‘low’ if they come in under 6-7% especially for RBs & WRs. QBs are a different story as they often are all clumped together.
- Dart Throw– Oh baby does this feel good when you hit it dead on. Dart throws are plays under two percent that honestly don’t have multiple paths to success. If they were an obvious play, they would garner more roster percentage. Hit one of these in a tournament and you are on your way to getting some cash. Also, expect to get these wrong on the regular.
- Leverage– This strategy is using players at either the same position (i.e. RB v. RB) or the same team AGAINST your opponents who are “popular plays”. It is also one of the best feelings in DFS and how you jump up the leaderboards. For example, 80+ percent of the time a Kyler Murray + DeAndre Hopkins stack is going to come through while shifting to James Conner instead doesn’t seem the most optimal. But if Conner falls into the end-zone twice, there is enough negative correlation between Arizona’s RB1 & Kyler last year to suggest you would sink those Kyler + Hopkins stacks. Gaining leverage allows you to create distance and smother all of those popular stacks.
Let’s Put This to the Test
Ok, we’ve mentioned roster percentages and how they work and I want to give you an example of what you could see in our Roster Percentage Report in the DFS Pass each week. In other words, the information you will have available to make decisions takes you that much closer to taking down a GPP.
Here’s a random example I pulled from last year in Week 8. The most popular RBs were as follows:
Now our roster percentage numbers will not be dead on but you would get a clear feel of who the most popular plays on the week and then begin to decipher why the field would like them. A couple of quick observations you could make:
- The two popular RBs are ‘backups’ whose salaries are affordable and allow you to pay up at other positions. They both are in weather games (CLE was rough that week & GB is Green Bay).
- Derrick Henry & Alvin Kamara were the two highest-priced RBs which makes sense given their track record. Henry’s opponent was begging us to play him.
- Myles Gaskin looked like the popular cheapie of the week given his workload and the perceived passing work he’d see against a favored Rams team.
- Jonathan Taylor had a great matchup against the lowly Lions and Jacobs was the other side of that Browns-Raiders weather game
- And all the way down at 8th that week was Dalvin Cook. The Vikings were 1-5 at the time and heavy underdogs against their division rivals on the road. While Cook arguably had just as high of a ceiling as Henry & Kamaram but with Jamaal Williams being the popular cheap RB, Cook’s roster percentage was deflated due to not wanting to roster two RBs in the same game.
Spoiler alert: Cook went bananas (50+ DK points) and broke the slate despite being only 10 percent rostered in the Milly Maker. Although we are giving some hindsight analysis, the key is looking at the information presented in the table and using the roster percentage projection to allow you to make more nuanced choices each week
So the question you can ask yourself each IF you have roster percentages presented:
- Who is popular and WHY?
- Based on this information, how would you attack this slate?
Let’s wrap up this look at roster percentages looking at the winning Milly Maker rosters from 2020 and our main takeaways in terms of salary and roster percentages.
As we’ve mentioned previously on the DFS Mistakes We’ve Made Podcast, leaving salary on the table is a sub-optimal move. But the average roster percentage gives you a ballpark to play around with. There are obvious outliers but shooting for somewhere between 100-125 percent is certainly doable. Take the projected percentages in the Roster Percentage Report and add your whole squad up.