Raw ratings are the number of votes that an option received in the poll. Calibrated ratings adjust the raw ratings based on individual voters' previous votes. For example, if someone has a history of overrating or underrating, their votes will be adjusted accordingly. Weighted ratings are applied on top of calibrated ratings to reduce the influence of outlier votes and simulate what the ratings would look like with more voters. Overall, this system aims to provide a more accurate and nuanced view of the opinions or preferences of the group being polled by taking into account individual voters' past behavior and reducing the influence of outlier votes. Does that make sense?