## loading packages used in project.
## setting conflict preferences.

What is this Analyis?

This notebook displays the estimated average, geek rating, and complexity for upcoming games. The model was trained on board games published through 2019, so I display the model’s estimates for games published from 2020 and onwards.

This analysis was run on 2021-12-30.

1 Estimating Ratings

We can first estimate the average rating and geek rating for all games given the data that is currently available on BGG. That means these sets of estimates make use of the current average weight for the game. This means the models are predicting BGG’s rating for a game given information we already from the community about the game. For games with relatively few votes, the estimate of a game’s average weight may be unreliable.

Additionally, we lose some of our ability to call our work predictive, in that the community’s estimate of the weight of a game is an extremely important indicator of how the community rates the game. To address these concerns, in the next section I first estimate the avgweight and then use the estimated value in the model.

1.1 Estimated Ratings for 2020

1.2 Estimated Ratings for 2021

1.3 Estimated Ratings for 2022

2 Estimating Weight and Ratings

These sets of estimates first estimate the complexity of a game, and then plug the estimated complexity into the model for predicting the ratings. This means these estimates will be relatively unchanging over time, with the only time varying factor being the year of publication.

2.1 Estimated Ratngs and Complexity for 2020

2.2 Estimated Ratings and Complexity for 2021

2.3 Estimated Ratings and Complexity for 2022

Get predictions for all.

Combine and save output.