Skip to content

Applying Portfolio Theory to Fantasy Football

With football season right around the corner, many fans are gearing up for another run at fantasy football dominance. Every fantasy league is full of players that use different strategies when drafting their teams, with some owners more willing to take on risk than others. Every league has a member that sticks to the safe, consistent picks, and someone that takes every injury liability and rookie on the board in efforts to strike gold. Whether you’re one of those players or somewhere in between, it is important to make the most of any risk you take on.

In order to optimize any risk taken on as a fantasy owner, one can use a version of a simple equation applied in portfolio management: the Sharpe ratio. In finance, the Sharpe ratio is defined as the average return of an asset divided by the standard deviation of the assets returns over time. A higher Sharpe ratio indicates that an asset is essentially providing higher returns for less risk. In fantasy football, a similar equation could indicate that a player would provide a high point total relative to his position group compared to the risk that would be taken on by drafting that player.

While it’s difficult to quantify risk in fantasy football, a look at the variation across different rankings experts would be a solid way to assess the uncertainty associated with a player.  In order to obtain this metric, I used the standard deviation of the 44 rankings used to calculate’s consensus rankings. This was used as the denominator in our modified Sharpe ratio. For the numerator, I found Z-score of each player’s total point projections (per relative to their position group. Points per reception (PPR) projections were used. This equation was applied to’s top 200 players list.

When interpreting the results below, it is important to keep a few things in mind:

  • A negative score does not necessarily mean that a player will hurt your team, but rather that he is projected to be below average relative to the universe of his position group that is ownable. Additionally, a player with a negative score will have a higher score with a lower ranking standard deviation based on the way the formula is set up, making it less applicable. However, a player with a lower score could be considered a sleeper candidate due to the fact that he’s a late round player with uncertainty among consensus rankings.
  • The scores are not useable to compare players from different position groups. Due to the slightly irregular distribution of some position groups, standard deviations aren’t perfectly transferrable between position groups.
  • Some players ranked at the top of their position groups were given with extremely high scores (Antonio Brown). Due to the nature of the equation, the marginal utility of the score diminishes quickly as it increases. This means that it can’t be interpreted directly against the rest of the player’s position group.

Below are the rankings with a short analysis, by position group.


Positive Scores


Analysis: Due to a consensus in the values of the top three QB’s on the board, they set themselves apart from the rest of the pack. Eli Manning provides above average return for low risk, making him a possible sleeper candidate for someone looking to wait on QB.

Negative Scores


Analysis: Tom Brady’s name on the bottom of the list can be disregarded,. He’s hurt based on the fact that his point total is only made up of 12 games due to his suspension from DeflateGate. With the highest standard deviation of ranking, Ryan Fitzpatrick represents a possible sleeper candidate despite his low performance projections. Philip Rivers represents a solid back-end starter with a low standard deviation and a top 12 ranking within the position.


Positive Scores


Analysis: David Johnson and Todd Gurley both come up as solid first round choices, with low standard deviations and high Sharpe ratio scores. Dion Lewis, Matt Forte, and LeSean McCoy all get a rankings boost from their low standard deviations, while Devonta Freeman represents a very risky pick based on his current 1st round grade. Similarly to Tom Brady, LeVeon Bell is unfairly penalized for the fact that he’s only going to be playing 12 games this year.

Negative Scores


Analysis: The low riskiness scores of Melvin Gordon and Chris Ivory suggest that they’re more likely to contribute as bench players, while the high riskiness scores of Wendell Smallwood and Devontae Booker suggest they could be good late-round flyers.


Positive Scores