It's easier to judge if you like or dislike something, rather than constructing a coherent personnal scale. It might be interesting to raise the cost of rating something, so you try to cater only to an elite audience, but you have less data. A simple 5 point scale is barely more difficult than a binary scale, I don't see the point very much.
Not only people have different tastes, but they have different scales. In most rating systems, moderates have less traction per unit than extremists, which means incentivizing the louder voices. The 1 or 5 situation in a 5 points scale system is just what's incentivized.
Different tastes, different scales AND different drive to rate things. Again, power to the clickers, apathy for the others. Bayesian can help here, but it is very limited. Regrouping and diluting ratings by user helps aswell, but is limited aswell due to the weight of anonymous ratings (if allowed).
All of this and the point of ratings, or critics, is just to have some good idea of what to expect. Rating here is meaningless without knowing what's the aggregated opinion of the average voter. The same way reading a critic review might be if you don't know the critic's tastes and style. Do your tastes match the average voter?
Paraphrasing homogenized, you are explaining you grew accustomed to some other ranking system, and you are used to a sweet spot elsewhere that you named "70-80". Answer is simple: different ranking system here, different sweet spot. You are just asking the ranking system here to match MU one's, without showing in what way it is better. You explain people are afraid to vote 3/5, which mean, by your logic, people are afraid to vote 6/10. Means 4/5 or 5/5 in a 5pts scale system, and 7-8-9-10 in a 10. Just adjust your sweetspot here. Why should the whole system be changed to accomodate people who have sweetspots in MU? What about those who have sweetspot in Batoto already? Should they adapt again?
Lastly, I'd be interested having a lot more numbers from the database, if that's possible Gendalph. Your computation is wrong, I suppose you did it by hand. I mean, "fol/view*1000" for Beelzebub is wrong at least. I am interested in the fol/vot ratio, wondering if the "rating" number is superflous, because redundant with "follow" (if the ratio follow/vote is fairly constant, at least relatively to the "5" votes).
Problem with Beelzebub is that I see there's a "popular comcis" list existing. Removes the need of "follow" for popular comics =(
- Chilled SDK likes this