Fantasy Football: Luck and Skill

Context: This work follows up on a previously written article about the relationship between luck and skill in fantasy football.

Objective: Use 2021 draft data to show that 'shared knowledge' has greatly increased the relative luck in Fantasy Football.


Previous Research

In our league, there is no relationship between time dedication and sucess. This seems to directly contrast the majority opinion that Fantasy Football is a "game of skill".


Our league showed no significant connection to historical win rates (Pearson Correlation of -0.19). The next step is to determine how shared knowledge plays a role in all of this.

Shared Knowledge

The chart above shows that preseason draft rankings are all very similar. However, they do a mediocre job of predicting actual performance. Moreso, this chart shows that no particular ranking seems better than any other.


This is what I am referring to as 'shared knowledge'. Different fantasy football sources all disseminate nearly identical information. So why do GMs stray from these sources and are they successful?


Research Hypothesis


Fantasy Football GMs that stray from rankings (demonstrating a belief in skill seperate from shared knowledge) will perform significantly worse than those that adhere to rankings

What does the data say?


From the chart above, one can see a positive correlation between a combined team FPPG and their similarity with Draft Reference Rankings (Spearman correlation of 0.66). With a p-value less than 0.02, this data would reject my null and indeed support my belief that shared knowledge greatly reduces relative skill in Fantasy Football


Note: The conclusion is consistent when comparing only starters and also when considering Total Fantasy Points instead of FFPG.

Conclusion

Null Hypothesis Not Rejected?


In this case, there was a significant difference, but it was the opposite of popular opinion. Shared knowledge proved better than any 'skillful' alternative


Research Limitations


Sample size was too small.


Our draft was held a month before the rankings were posted.


Our league has slightly different rules and point calculations than the rankings assume. For more on this, see my work on Defensive Schemes


Next Steps


It would be nice to build a website similar to this one that allows people to analyze their league's draft data while also feeding a database.

Don't go cursing off Fantasy Football just yet...


Admittedly, I lied with the data. All the numbers and labels are accurate, but the conclusion is misleading. Can you figure out what it is? Feel free to contact me if you would like to know more about this research topic and what I have done beyond what you see above.