Another list that makes no sense since it is highly favoring lesser played titles which would receive lesser negative reviews. I mean games like Valheim, Inscryption, GoTG, etc all have OP reviews and some def have 10's of thousands+ of reviews. But whatever.
It definitely is not sales and player count list, it seems to punish titles with larger review/audience numbers more than it should (they have a link to the algorithm they used on the page if you want to see what they did). Like number 11 has 3415 positive reviews and 28 negative with a peak concurrent of 412, While 2 below it is valheim with 299,548 positive reviews, 13,251 negative and concurrent player counts of 502,387. Would number 11 still be rated what it is if it had ~300k reviews?
From what i read, their algorithm will "punish" titles with less reviews.
They use a certain estimate of uncertainty that decreases at a logarithmic proportion, its 100% uncertain at 1 review, 50% at 10 reviews, 25% at 100 reviews, 12,5% at 1000 reviews and so on. The calculation goes as such.
If game X with 90 score has between 10 and 99 reviews (that would be 50% uncertain), the rating will be defined by:
90*0,5 + 50*0,5 (where 50 is considered as the absolute average score)
So the rating of this game will be
70
If game Y with 80 score has between 1000 - 9999 reviews, that would be 12,5% uncertainty. So:
80*0,875 + 50*0,125
So the rating of this game will be
76.25
In other words, a popular game with a 80 score would still place above an obscure game with 90 score.
The difference starts becoming somewhat negligent after 1000 reviews, but at that point i think its safe to say the pool of reviews is big enough to bring about a fair scoring.