NFL DFS Strategy: 2021 DraftKings Scoring Analysis

Jon Anderson breaks down how various positions have scored on DraftKings and provides some takeaways on what it means for how you should build your DFS lineups in 2021.

We talk a lot about NFL DFS here at RotoBaller. It is a wildly popular game and in fact, it proves to be much more strategic and beatable than season-long fantasy football. By that I mean that the best players will win at a higher rate in DFS than the best players will win in season-long fantasy football.

The weekly selection salary cap model makes the game much more nuanced, with many more angles to attack and a clean slate every week where an injury can only hurt you for one week rather than many. You see a bunch of DFS articles being posted here throughout the week as we prepare you for the weekend. Most of those articles will focus on individual slate to come. This article is DFS-centered, but it is a bit different. I am taking a high-level look at what has happened in the first five weeks of the NFL season.

The way I've done this is by collecting a data set of every player's box scores alongside the game information (like spreads and totals), and their DraftKings salary and scoring data. This dataset proves to be very powerful in generating overall insights. Let's go through it.

First, let's take a look at how many DraftKings points each of the league's defenses have allowed. To do this, I summed up all quarterback, running back, wide receiver, and tight end DraftKings totals from the five weeks we've played. Here's the data, you can sort by any column you want.

We are still quite early in the year, so the schedule plays a major factor here. The Dolphins have allowed the most fantasy points, but that's mostly thanks to the 173 they just gave up to the Bucs in week five. Only five teams have had to play that pass-heavy offense, so it's not really fair to judge defenses in this manner so early in the season. Most of these top teams have played tough schedules so far, and the teams near the bottom have typically played soft schedules. Don't take this table overly seriously, but we shouldn't be completely ignoring it either.

Now let's quickly check DraftKings scoring by team and position.

Here we see the Bucs leading the way by a good margin, 38 DraftKings points ahead of the Chiefs. The common denominator in the teams at the top of the list is elite quarterback play. You have to go the whole way down to the 9th team on the list (Vikings) to find a quarterback that wouldn't widely be considered a great fantasy option. At the bottom of the list, we see the young, inexperienced quarterback play with the Bears, Jets, Dolphins, Texans, Jaguars, and Patriots occupying seven of the bottom seven slots. All of those teams are either starting a rookie or have had significant injuries at the QB position. You can parse through that table at your leisure.

Next, I wanted to quickly look at average DraftKings scoring by position. For this, I used just "starters", which I considered as players that saw 30 or more snaps in a game. That way, we don't have backups bringing down the averages and skewing things. Here are the results:

Despite the league's shift to a pass-heavy league once again, running backs are still outscoring wide receivers on average. This has probably more to do with only a couple running backs on each team getting to that snaps played threshold, while four or five wide receivers might play 30+ snaps in a game and they simply all can't score a bunch of fantasy points. This doesn't tell us that much, but it's interesting to see how the positions compare. You really do want to get a strong score out of your quarterback when building a DFS lineup, because you can be sure that a lot of the field is going to grab 20+ points from their guy.

Now let's look at some more interesting stuff. For each position, I looked at every fantasy-relevant (using that 30 snap threshold again) and grouped them together by their DraftKings salary. I binned them all off by rounding their salary to the nearest $500 mark and then averaged out each groupings scoring. Here is the QB data:

We can see that as price increases, so does average scoring. This is a good indicator that the DraftKings pricing algorithm is pretty darn sharp. The reason we see the most expensive group taking a step backward is just a lack of data. Only two quarterbacks have reached the $8,500 pricing bin (Patrick Mahomes in week two, Kyler Murray in week three), scoring 28 and 23 DraftKings points respectively.

One interesting thing from here is the $5,500 and $6,000 scoring being very similar. This is not a sample size thing either, as there are 41 quarterbacks in the $5,500 group and 28 in the $6,000 group. At this point in the season, it would seem to make some more sense to favor either the very expensive quarterbacks or the cheapest viable options. That said, this is high-level analysis, and when you're actually picking a quarterback for an individual week the granular details are very important. Now let's check running backs.

Here we see the same general trend with points increasing as the price grows. The $8,000 bucket you see there with a lower than expected total had a lot to do with the Saquon Barkley injury last week. Only six running backs are in that bin, so Barkley's 3.7 point performance really pulled down that average. We also see the $5,500 and $6,000 categories bolstered quite a bit by Alexander Mattison's two big games while stepping into the starting role with Dalvin Cook on the shelf.

Those two two price bins are both just single Christian McCaffrey games, and the bins at $8,500 and above are all only including CMC along with Alvin Kamara, Dalvin Cook, and Derrick Henry. The majority of running back prices have fallen between the $5000 and $7,000 buckets. Overall we do see expected growth here, there are not any really good takeaways here. With several teams splitting up carries between their running backs, we've seen how valuable paying up for the big names can be. Getting those 25+ point games from your RB slots really gives you an advantage since they are quite few and far between this year.

Now wide receivers:

These bins are much more evenly distributed in terms of how many receivers are under each one. The top two bins have just Davante Adams (once in the $8,500 group) and Tyreek Hill (once in the $8,500 and once in the $9,000) in there, which is why you see the smaller averages there.

The $8,000 bin is the winner, averaging 22.8 DraftKings points. That bin includes Adams, Hill, Calvin Ridley, Cooper Kupp, and DeAndre Hopkins. Ridley and Hopkins have really disappointed relative to cost in those games, but we also saw a complete smash game from Tyreek which boosted that number.

The salary bins with significant numbers of wide receivers in them are everything between $3,000 and $7,000. Between those two ranges, you see growth in points scored, with a huge jump from $5,500 to $6,000. There are a lot of big-time duds at the $5,500 level with much fewer at the $6,000 level. There also appears to be more ceiling at this $6,000 level with guys like Kupp, Deebo Samuel, Terry McLaurin, Mike Evans, and Diontae Johnson all going for 24+ points while priced at that level. We have seen only two performances of 24+ points at the $5,000 level  (and one of those was the super weird Cordarrelle Patterson game), and just one at the $4,500 level (Christian Kirk). The $4,000 price range average score of 8.7 is low, but it should be even lower since the Kadarius Toney game when he was forced into targets by multiple injuries is being factored into that.

The point is, pass catchers are tough to rely on at the lower price levels, and that is especially so in situations where injuries aren't playing a factor. Most weeks, there should be injury situations to exploit and that's a good time to look into these sub-$5,000 wide receivers to save some salary because we see that wideouts really can post some big scores when more targets become available.

Now let's finish this section out with tight ends.

This is a different situation than the rest of the positions because there just aren't that many tight ends getting fantasy points. We have seen incredible low-scoring outputs for the cheapest tight ends, and we haven't gotten much from the mid-range guys either. The $5,000 tight ends have averaged just 10.3 points while the $4,000 guys are less than 1.5 points behind. The top of the line tight ends (all Travis Kelce and Darren Waller), have indeed been much more reliable, but even then you aren't seeing must-have scores coming from the tight end position (the final bin there is just one game from Kelce).

We should, of course, take every slate on its own and consider each circumstance, but from the high-level view, it seems the way to go is just to look to punt tight end so you can have money for the elite players.

Correlations

A huge part of the DFS game, especially in tournaments, is exploiting correlation. Every passing play involves more than one player, meaning that quarterback scores will end up having some correlation with other positions. I checked on this correlation with a scatter plot:

For each team, I looked at each game they've played and summed up QB and WR points, and then plotted them all together with QB points on the x-axis and WR on the y-axis. You can hover over each dot to see what team and week the dot represents. So the dot way up top at the right side of the plot is Tampa Bay's week five-game, where Tom Brady scored 42 DraftKings points and his receivers summed up to almost 100. In the bottom left, you see that putrid Texans game where Davis Mills went for -1 points and his receivers, obviously, were held under 15 points.

The trend is upwards, which proves that the correlation does exist. This isn't a surprise. What I take from this is that the re-emergence of the running quarterback is lowering this correlation. Any dot well below the line is an "outlier" dot, which isn't following the trend. If you hover over those, you see teams with mobile quarterbacks (Ravens, Bills), as well as a lot of teams that throw to running backs a bunch (Washington, Cleveland, Detroit, New Orleans). This year more than ever, quarterbacks can score big points without their receivers following suit. You absolutely do not need to play wide receiver with your quarterback if you're playing a running quarterback.

I ran another correlation on quarterbacks vs. running backs.

We would not expect much of a correlation here, and indeed we don't have much of one. The fact that QBs throw a decent amount of touchdowns to RBs makes that line slightly trending upwards, but overall running back scoring is pretty disconnected from QB scoring. That means it doesn't make a ton of sense to play a QB along with the primary RB.

The last player correlation I did was QB with the WR1 on the team. I defined WR1 as the wide receiver that has scored the most fantasy points so far this year for their team. That means that I'm not using the actual best receivers on the teams, but I thought this was the better way to go for the purposes of this analysis. Here's the plot.

We see a slightly stronger correlation here. When a QB goes over 30 points, we only see four instances of the WR1 being under 20 points, while we see almost 20 instances where the WR1 went over 20, and there are about ten instances where they went over 30. This is a little bit unfair given the way I did the analysis, because I'm only using wide receivers that have already led their team in scoring, but it proves a pretty good point nonetheless. If you're playing a quarterback that scores most of their points through the air, it's a pretty good idea to pair them with their best receiver.

One last correlation I checked was DraftKings scoring as compared to the game totals. We typically favor game environments where lots of points are expected to be scored, which of course makes sense. It is worth checking on this before just blindly accepting it, so here's the chart:

Another upward trend here proving there is some correlation present. We do see a large number of data points hovering quite far away from the line, proving this is by no means it's a sure-fire thing that high totals will result in high fantasy scores. The highest two games in DraftKings scoring this year have had totals below 47. We saw a game with a 55 point total go for a very average amount of DraftKings points. While the general trend is upwards, there are much more important inputs than game over/under when trying to predict fantasy scoring.

Conclusion

This post was more descriptive than prescriptive, as we didn't find a ton of stuff that we wouldn't have expected. But here are some takeaways that may help your lineup building moving forward:

• A general rule of paying down for tight ends seems like a good idea given the lack of proven scorers at that position.
• The elite running backs are very valuable given the scarcity of the guys you can lock in for 25+ points.
• Do not feel pressure to pair a running QB with one of their wide receivers.
• The game total does not correlate as well as you may have thought with expected DraftKings scoring.

Download Our Free News & Alerts Mobile App

Like what you see? Download our updated fantasy football app for iPhone and Android with 24x7 player news, injury alerts, rankings, starts/sits & more. All free!

More Fantasy Football Analysis