Game Of The Year Nominology

About the Predictive Model

By: Chief Game of the Year Nominologist Kaiser Kyle

Below are the most recent results of our Game of the Year nomination machine learning model. This model uses MetaCritic review data and other basic information about each game to attempt to predict the most likely nominees for Game of the Year. Included in the data set are things such as the raw number of critic/user reviews, the average score of critic/user reviews, the ESRB rating, release month, and others. The raw probabilities are in the graph on the right. These can be interpreted as "In an average year, what percent chance would each game have at being nominted?". The graph on the left simluates what would happen if the nominations were today by scaling the results so that the sum of all probabilities equals 500%, as there are 5 nominess each year.

Notes For Consideration:

• Only games with a critic score above 70 and over 10 user reviews are considered
• The model predicts for all valid games released this year, meaning that there are many more games not listed which are at less then 1% chance, so the numbers you see in the left graph do not add up to 500%
• Since the model uses the raw number of critic/user reviews, there is a ramp up time as more and more reviews come in. Newly released games will not immediately rise to the top

Most Recent Results:

Last Updated:

Raw Data

Note: A "0" in the ESRB Rating column means unrated.

Changes Since Last Update

Below you can find the changes in scores and review counts between and . Games that are new to the list this week will have no data, as they were not present in the previous week's data.