It’s the same system that IMDB and RateYourMusic and many other rating driven websites use.
The premise is twofold: (1) as the number of ratings increases, the rating on the game better approximates the true rating, (2) as the number of ratings increases, the rating on the game will decrease.
(1) is true for obvious reasons. Consider a case of a story rated 8.0 with 1 rating, and a story rated 7.7 with 1000 ratings. Which one are you more confident about? Obviously the second story, even though the first story has a higher rating.
(2) is often true because as a game gets widespread attention, it’s likely being played by fans outside the genre and by a more diverse caste of people. In that case, it’s harder to maintain a high rating. While this is often true, (1) alone is sufficient to demonstrate the utility of the system.
At a certain (relatively low compared to Endmaster number of ratings) number of ratings, adjusted rating basically equals true rating. Go hover over some games with a lot of ratings and see how close those two numbers are. If the top games aren’t dislodged, it’s not a problem with the rating system, it’s a problem with the ratings.
Although, it would be an interesting thought experiment to also diminish the rating weight on older ratings, to account for the gradual increase in story quality.