Skip to content

Create statistical model behind the rankings #8

Description

@debruijn

With my data background, I can't resist to think about more advanced things to do with this data than just these "simple" aggregations and transformations.

Downside: the simplicity can be a good thing since it is easy to explain. The most difficult step in the repository right now is the transformation to bring the points of the rankings in line with each other, and even that is easily tractable. Creating a statistical model will lead to more decisions, more room for debate and disagreements, and more explaining the results.

Upside: modeling is interesting and can lead to new insights. We could look into a hierarchical model on the (transformed) scores per team per ranking, and use the higher-level location parameter to make statements about how likely it is that team A is better than team B given the points of both teams for all three rankings. Also, we can impute missing ranking values for rosters that are not tracked by all three services.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions