Edit 1: This post has been edited to include a couple more scoring settings, a better explanation of the method, and an important limitation.
Edit 2: I have taken the data and built it into a league comparison web app so you can directly compare 2 leagues, including seeing how many points the position will score and who the top players would have been in 2020. You can access it here:
Edit 3: part 2 is now out focusing on how league settings affect draft strategies
Every year, Fantasy commissioners look to tweak their league settings to try to make the league fairer, more fun, or better match the NFL. Often these changes are done with a vague understanding of their effects. Does moving passing TDs from 4 points to 6 points make QBs more valuable? By how much? What about half-point vs full-point PPR? We all know to look for rankings that resemble our league settings but sometimes these are not available. And when we are deciding on league settings, how are we supposed to know which settings to tweak to have the desired effect.
In this 2 part series, I will break down how a few common league settings affect positional value. In this post, I will focus on how the settings affect positional value at the end of a season. In part 2 I will focus on what different league settings mean for different draft strategies.
The Data
I will be using data from 2007 to 2020 with fantasy points calculated based on pro football reference stats and ADP data (for part 2) provided by the fantasy calculator. I will also be ignoring defenses and kickers for this post. The focus will be on the value of positional players.
General Explanation of the Analysis
For all league settings, I will be comparing the positional breakdown of the 40 most valuable players in the league. In order to calculate player value, we will be using value over replacement player (VORP).
Value over replacement is defined as the number of fantasy points a player scored minus the best non starting player at that position (so in a 10 team league, this would be the points over the 11th best QB or 21st best RB assuming 1 QB, 2 RB league). I will also be looking at different numbers and types of flex positions, please note that for all of these, I will first fill the starting positions before looking at the players available to fill the flex position(s). Vorp will also be calculated for the flex position as if it were a unique position. For example, if we are in a 2 RB 1 flex 10 team league, the replacement player for RB will be the 21st RB, even though that RB will likely be starting in the flex position. I will be playing with the number of starters so this point will change but know it is always the number of points above the best non-starter. Also, note that VORP has a minimum of 0, no player will have a negative VORP.
Once I have VORP calculated, I will select the 40 players with the highest VORP in each league setting option for each year. I can then just count the number of QBs, RBs, WRs, and TEs who are on that list. While not perfect, this does give a sense of how positional value changes based on the settings.
For all graphs below, I calculated every combination of the settings I looked at for each year (over 300,000 unique combinations). For the graphs, I will include all data points and just split them by whatever parameters we are looking at (or are mentioned). All combinations of all other parameters will be included and should therefore cancel out, but will provide a nice sample of the possible effects when we combine different parameter choices.
We can start by looking at positional value year to year in a classic league format:
(1 qb, 2 rb, 2wr, 1 te, 1 flex)
We can see that RBs and WRs are the most valuable (and surprisingly equally valuable). While QBs and TEs are on a separate tier of value. Additionally, this value does not have a consistent change year to year, at least not since 2013ish.
Ok, given these let's start to compare some settings
8 vs 10 vs 12 teams
The changes are small for the most part, but QBs value increases as the number of teams increases while TE value decreases. Running back and Wide Receiver value remains constant.
4 Point vs 6 Point Passing Touchdowns
Changing the value of passing TDs doesn't change much in terms of positional value, though QB value is increased slightly. The WR value decreases slightly to make up for the difference. The larger effect here is likely which QBs are valuable, rather than making QBs more valuable when compared to other positions.
Passing Yard Value
Another common strategy is to change the value of passing yards. Similar to changing the value of passing touchdowns the effect here is limited particularly to the extremes. The same as with passing TDs, this is less about changing the value of the QB position as a whole and more about changing which QBs are valuable.
What about PPR?
Unsurprisingly, running backs are most valuable in 0 PPR, while WRs are most valuable in full PPR. Tight ends are also most valuable in full PPR while QBs are the most valuable in 0 PPR.
What about a TE premium?
The idea of a TE premium is to add extra value to TE receptions above whatever the normal PPR is. So if this is a full point PPR and the TE premium is .5 points, TEs get 1.5 points per reception. Unsurprisingly, this changes the value of TEs, roughly so that they are in line with QBs.
Ok enough changing scoring rules, what about changing roster construction?
1 vs 2 QB leagues
Unsurprisingly, increasing the number of starting QBs significantly increases QB value (about doubling it) while all the other positions take a hit to make up for the difference.
2 vs 3 Starting WRs
Similar to adding a starting QB slot, adding a third starting WR spot significantly increases the value of WRs while decreasing the value of all other positions.
What about adding a Superflex?
A common choice to try to increase QB value is to add a Superflex position (a starting position for QB, RB, WR, or TE). Adding a Superflex position seems to have a similar, though less extreme, effect as adding another starting QB position.
Limitations
The main limitation here is that all of this analysis is on season-long results. The positional value will change year to year and week to week. It also doesn't take into account things like QB streaming or other strategies that affect positional value. I also did not look at the value of specific players which settings like PPR and passing TD value will certainly affect. That being said, analysis like this provides a sense of how value changes based on league setting.
Another limitation is that there are other factors that will affect the positional value that I do not cover here. For instance the drop off if a player were to get hurt with different positions being more likely to be hurt compared to others (which was some of the original arguments for zero RB draft strategies).
Finally, I am hesitant to use this to assume positional value overall (regardless of league settings). Not only do league settings change this, but many of the settings I am looking at are also specifically geared to impact QB or WR value. As such, the averages I present are likely biased towards these positions. This does not affect the take-home points about how each individual setting impacts positional value (since all combinations are included in all groups), but I hesitate to generalize the results beyond the singular comparison. If you want to see the value for a specific combination, try my new web app here: https://ffleaguecomparison.anvil.app/
Conclusions
Overall, this was an interesting exercise. While the trends are basically what I expected, but it is cool to see actual numbers attached to the effects. Remember to take these into account when joining a new league or when changing settings. I want to be clear that I think there are reasons to choose any combination of these settings, which boils down to personal preference. The goal of this post is now we can see what effect these settings have so we can make informed decisions relative to each league's goals.
In part 2, we will look at how these changes in positional value affect draft strategy.
All code related to this post is available on my github.
Questions? Comments? Let me know at ac@alexcates.com. Want to read more breakdowns like this? sign up for my newsletter. Finally, like what I do? Consider supporting me on buy me a coffee.
Comments