I’m sure this has been discussed before but I can’t find it off the top of my head…
On the map list there’s a “Users” column for the user rating. On the map page itself there’s also a “User Rating”. These two are often different.
I thought I vaguely remembered that one of these was using a weighting function while the other was a raw average?
Although if you look at something like https://www.quaddicted.com/reviews/arenaq.html … on the maps list it has a Users value of 2.16, while the rating and histogram on its individual page seem to show that everyone rated it 1.
So. What’s up?
Some info is at the bottom of https://www.quaddicted.com/help/maps
Maybe I should hide the weighted rating, people have whined in the past. It was mostly meant for sorting. Theoretically one could fit it into 1-5 also but I wonder if that would be even more confusing?
Nah Spirit, everything is good as it is.
Imagine, a good map gets one single rating from a funny clown: 1
It would be at the bottom of the list when using the User rating filter.
Or the other way round, if a bad map gets one single user rating: 5
It would be at the top of the list. → Best map ever
That is why Spirit uses a rating system that starts in the middle and slowly climbs up or downwards, depending on the many user ratings that are about to come.
It is the only fair way to prevent misuse and after around 10-15 votes the rating is where it belongs.
The more people vote the more exact it will be.
It also prevents ratings with only a few votes from jumping up and down in the list.
If people want to see the exact average, they will find this value on the maps page.
Thank you Spirit for your work at Quaddicted. A great place to be and a perfect place for Quakers!
People will not stop asking this question and we will not stop answering it.
Thanks Spirit, makes sense. Might be nice to also see the Bayesian average on the maps page itself, to emphasize that yeah this is a different thing from the raw average and not a mistake. But whatevs.
Thanks! I think we all agree that the weighting is important, but the numbers are confusing indeed.
Try https://www.quaddicted.com/reviews/index_normalized.php for how it could be like instead. Better?
I admit I’m not sure what I’m lookin’ at there.
Let me stick with arenaq as the example. On its map page, the rating and histogram seem to show that 10 people each rated it 1.
The “normal” reviews list shows its Bayesian average rating as 2.16.
The alternate list linked above shows it as 1.33. Just a different weighting?
https://www.quaddicted.com/reviews/index_normalized.php is the bayesian averages made to fit the 1-5 range of ratings for making it feel better.
Ahhhh ok. Yeah those numbers are a little less surprising in the case of the low-number-of-votes outliers.
Don’t have any proper arguments for it, but the normalized version feels better in my book.