The average rating for projects that receive a low number of ratings will have very high variance and poorly represent how participants felt about your game, so we do not show a ranking for that score. It would be unfair.
How is that unfair to just show the information? I don't think it's fairer to negatively compensate dispersion, dispersion works both ways, and in many cases, you would lower a score that was already too low from not enough ratings.
That might be worth adding, but in this case you would still have too low number of ratings to make a difference based on how we would recommend hosts to configure the jam.
I never said that my 8 ratings should be enough, but I should at least get some warning that I should try to get more ratings and how many more. Will you add this feature?
10th percentile would be way too lenient, and allow for games with high variance in their rating overtaking projects that have a more accurate rating due to the larger number of ratings.
The system is built this way to encourage you to participate in the rating process. If you want to receive ratings to be eligible for a more accurate and non-penalized final score then you will need to rate other people’s work and leave constructive comments on their submission pages. That will allow your project to show up on the “most karma” sort and also people who see your comment will be linked submission to get a chance to play and rate your project back.
I see the intention, but in this particular solution, wouldn't you punish half the people even if everyone got 100 or more ratings? Median will always shift if more people get more ratings and there seem to be no way to get zero people penalized. Also, even if I do get enough ratings to pass the median, I still know that half the people got their scores lowered and my game is compared to them not objectively.
I still think that Ludum Dare has the best solution for this problem. Everyone can get 20 ratings and they are forced to play and rate games of other people. Also, everyone gets a very clear warning that if they have lower than 20 ratings, they will not be ranked. In their case, every participant can get a fair score and be ranked among others, not just half the participants. I see no reason to leave this median solution, it creates more problems than it solves. You literally make really interesting rating data lose its objectivity :(