DFS: How Well Does Quarterback Salary Predict Points? (Fantasy Football)
The most important thing to consider when building a DFS lineup is player salary. A player’s price point is a lot like a betting line. Most of the information that you are taking into account in your own research is already baked into this price. If you want to have any chance of making money over the long haul in DFS, you have to be very sensitive to and smart about salaries.
I have collected a huge data set of the last six NFL seasons. In this data set, we have every single offensive player’s stat line along with their DraftKings salary and points total for every individual game.
QB Salary History
One interesting thing to note is that DraftKings has altered their pricing for quarterbacks over the years. I looked at all QB’s who played the majority of their team’s snaps for each game and checked their average salary for each of the last six years, here are the results:
DraftKings was pretty aggressive on pricing back in 2014, but they have been pretty consistent since then. We will restrict the rest of this post to the last three seasons since QB salaries have looked pretty much exactly the same since 2016.
Salary vs. Points Relationship
The question of the day is how good of a job the algorithm does at pricing quarterbacks. If it was a perfect algorithm, you would see the average QB points scored increasing incrementally with every $100 added to salary. Let’s take a look to see what it really looks like.
Here is just a scatter plot of all QB performances over the last three years.
Each dot here represents one QB in one week of play. For example, the rightmost point you see there is Tom Brady in week six of the 2017 season. He was priced at $8,300 and scored 17 DraftKings points that week.
The highest point on the plot is Aaron Rodgers in week seven of the 2019 season. He was priced at $6,400 in this one and he torched the Raiders for 429 yards and five touchdowns.
What scatter plots do a great job is showing the trend. Here, you see a general trend upwards – as salary increases, so do points. However, it is clearly no guarantee to happen.
Next, I took the average QB DraftKings score at each price points and plotted it. Here’s what that looks like:
You can see the upward trend much more defined here. The ends of the plot (the lowest and highest price points) are a bit all over the place, but this is mainly due to there not being many data points there.
For example, you see that quarterbacks priced at $4,000 average around 15 DraftKings points, while the $4,100 and $4,200 players do much worse despite being more expensive. There are only 11 players priced at $4,000 and just seven priced at $4,100 in the sample, for example – so one game has a big impact. The same is true at the most expensive price points. There is just one QB priced at $8,300 in the sample, one at $8,100, three at $8,000, one at $7,900, and so on until $7,500 when there is actually a large amount of data points there. So yes, this plot does suggest that $7,900 quarterbacks are better plays than $8,300 – but that is just because there was only one instance to look at.
The salaries between $5,000 and $7,000 is where we can really glean some insight. You can see that once you get past $5,700 or so – you really aren’t gaining much by paying more for a quarterback. The average score for a $6,000 QB is 19.5 points. The average for a $7,000 QB is 20.5. There is no noticeable difference between those price points.
Next I went one step further and rounded every salary to the nearest $500. This gives us 10 salary “buckets” to look at it, from $4,000 to $8,500 counting every $500. Here is the plot:
Once again, that pesky dot on the far right is just that one game from Tom Brady, he was the only QB of the last three years to be priced at $8,300 or above, so he is alone in the $8,500 bucket.
The rest of the data shows the same as what we saw in the last plot. You see a sharp rise every $500 until about $6,000 and then things even out until you get to $7,500. A $6,000 QB and a $7,000 QB are separated by just 1.5 points. Here is that plot in table format:
The DraftKings pricing algorithm does a pretty good job. The highest priced quarterbacks will score the most points. However, there does seem to be a salary range we can exploit. The data for the last three seasons shows a pretty negligible difference between quarterbacks priced in the $6,000-$7,000 range. It may be a profitable strategy to stick to the $6,000 quarterback most weeks and save yourself some money to use on other positions.
We will go through similar studies with running backs and wide receivers/tight ends in the coming weeks, so keep on coming back!
Whether you’re new to daily fantasy football or a seasoned professional, be sure to check out our Daily Fantasy Football Glossary. You can get started with A Beginner’s Guide to Daily Fantasy Football or head to more advanced strategy — like Maximizing Your Potential in Multi-Lineup Contests — to learn more.