ynot wrote:I've been digging into the whole question of "do matchups matter?". I think it's of interest with the season coming to a close which means strength of remaining schedule is becoming more variant, especially with the competition gap. It might also be helpful for H2H players looking to target specific weeks.
Here are some charts looking at the 2021 season for players playing over 40 minutes, to filter out injuries and small role bench players.
This looks at the correlation between game margin and total fantasy points scored. Margin being used as a proxy for the difference in how good a team is in a matchup.
The R = 0.25 refers to the correlation between the two variables, 0 being no relationship, 1 being a positive relationship.
This means there is a positive but weak relationship, between individual fantasy points and team margin. i.e. The more points a teams wins by the more fantasy points an individual in that team will score.
I think this is fairly obvious, because the events that occur when scoring game points also create fantasy points.
Lets see what happens when we break that down by position:
I think this also makes sense, backs are the most effected by game results (matchups), particularly fullbacks and halfbacks both having moderate positive correlations. Centres to a slightly less extent and Five Eighths even less.
Forwards on the other hand it doesn't matter so much, I'm guessing because their fantasy points are generated by tackles and run metres, which remain relatively consistent despite opposition.
What doesn't remain consistent between players is minutes played, so lets look at the previous chart on a points per minute basis:
A slight increased in correlation but nothing major.
To summarise: trade in backs, particularly halves and fullbacks, who have good matchups and don't trade in those that have tough matchups. Don't worry too much about the matchups for forwards!
Pretty obvious, but I'll hopefully dig into this a little deeper though out the week and come up with some recommendations.