ubuntu-phone team mailing list archive

[Matthew Paul Thomas <mpt@xxxxxxxxxxxxx>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Martin Albisetti wrote on 09/04/15 13:38:
> ...
> 
> There are a lot of good points about people gaming the system and 
> problems with different approaches. Luckily, this is not a new 
> problem at all. In fact, it's not even new to Ubuntu!  Matthew
> Paul Thomas spent some time thinking about this years ago and put 
> together a great wiki page[1]. Which points to a great post on how 
> to avoid the most common problems: 
> http://www.evanmiller.org/how-not-to-sort-by-average-rating.html
> 
> ...

The sorting algorithm we ended up with is described in the main USC
specification. <https://wiki.ubuntu.com/SoftwareCenter#top-rated>

"An item’s dampened rating (DR) should be 3 plus the sum, for each
possible number of stars (1, 2, 3, 4, and 5), of (that number of stars
– 3) multiplied by the lower bound of the Wilson score confidence
interval with a 90% confidence level for the proportion of ratings
that had that number of stars."

This is similar to Selene's suggestion. It avoids messing with the
displayed rating -- an item with only two reviews, both 5-star, will
show an average rating of 5 stars, so people don't think the average
is broken. But that item will sort below an item that has nine 5-star
ratings and one 4-star rating, even though the raw average is higher.
This is more understandable, because sorting can also be influenced by
other factors as Martin mentioned -- popularity, usage, crashiness,
serendipity, and so on.

- -- 
mpt
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iEYEARECAAYFAlUrlDkACgkQ6PUxNfU6ecoQRACeIRyrJFa+MvUD0DquPNJ4ptZq
8VkAnitPUgmTjvWO4KqOiC5mnQDGKVQN
=Y8mV
-----END PGP SIGNATURE-----