By: Suvrath Penmetcha (@savvysuv)
With the Fantasy Football season upon us in just a couple days, we wanted to take a closer look at year-to-year consistency among QBs across statistics (i.e. TDs) to learn how past performance can be applied when drafting players.
To that end, we used the nflscrapR package and the season_player_game() function to find the statistics for each player for each game. We were then able to calculate per-game statistics for QBs, for each season from 2009-2015, across key attributes which contribute to QB scoring in standard fantasy leagues. These include
- Pass Yards
- Pass TDs
- Passing 2-Point Conversions
- Rush Yards
- Rush TDs
- Rushing 2-Point Conversions
We were then able to calculate the correlations for each of these key attributes across seasons.
The visualization above is a correlation heatmap, which identifies the correlations between a quarterback’s key scoring attributes across seasons. Using this information, we can highlight the most significant relationships and draw several observations.
- Fantasy points in Year 1 and fantasy points in Year 2 are strongly associated, with a correlation coefficient of 0.9; this means that high-performing fantasy QBs have a tendency to replicate their success the next season.
- The most repeatable attribute among quarterback performance is rushing yards, followed by passing touchdowns and passing yards. Based on these results, QBs who had a lot of rushing yards and passing touchdowns last season should be prioritized, because there is a strong chance they will do that this season as well.
- We can also use this heatmap to identify see which attributes aren’t as repeatable from year to year. For example, one of our most interesting findings is that there is a low correlation between QB rushing touchdowns across seasons. So, it is important to note that quarterbacks are generally inconsistent about getting into the endzone.
Overall, this graphic helps us to see the repeatability of key attributes that contribute to QB fantasy scoring and it can help determine new draft strategies as they relate to quarterbacks. In future articles in this series, we’ll repeat this analysis for other positions such as running backs, wide receivers, and tight ends.