Print Page | Close Window

Strange Change in Album Rating

Printed From: Progarchives.com
Category: Site News, Newbies, Help and Improvements
Forum Name: Report bugs here
Forum Description: Help us improve the site from a tech standpoint
URL: http://www.progarchives.com/forum/forum_posts.asp?TID=51634
Printed Date: March 03 2025 at 21:55
Software Version: Web Wiz Forums 11.01 - http://www.webwizforums.com


Topic: Strange Change in Album Rating
Posted By: Roj
Subject: Strange Change in Album Rating
Date Posted: September 08 2008 at 09:23
I hope I am posting this question in the correct section. Here goes, anyway. 
 
Today I placed a review of Chronometree by Glass Hammer, which before my review was averaging 2.92, based on 27 reviews.  I gave the album 4 stars in my review, yet it's average has now gone DOWN to 2.89 following my review!  Any ideas why this has happened?
 
Thanks, Roj.



Replies:
Posted By: chopper
Date Posted: September 08 2008 at 10:16
It may be something to do with collab reviews getting a higher rating, but you really need MikeEnRegalia to explain the algorithm.


Posted By: MikeEnRegalia
Date Posted: September 08 2008 at 10:54
^ I used to know how it works ... but M@x made some significant chances since then.Embarrassed

-------------
https://awesomeprog.com/release-polls/pa" rel="nofollow - Release Polls

Listened to:


Posted By: Padraic
Date Posted: September 08 2008 at 11:00
Originally posted by chopper chopper wrote:

It may be something to do with collab reviews getting a higher rating, but you really need MikeEnRegalia to explain the algorithm.


That should only mean that the overall rating wouldn't increase as much as it would if a collab gave it four stars....but there's no explanation for the rating decreasing in value.


Posted By: chopper
Date Posted: September 08 2008 at 11:16
It would if the algorithm caused Roj's rating to be less than the 2.92 average. I may be barking up the wrong tree anyway, looks like we need mailto:M@x - M@x to explain.


Posted By: Roj
Date Posted: September 08 2008 at 11:23
Excuse my ignorance folks, but do I need to e-mail mailto:M@x - M@x or post something elsewhere, if so I don't know how. If mailto:M@x - M@x will see this and help out, that's fine.
 
A system whereby a positive review can actually cause an album's rating to fall is clearly not ideal.


Posted By: Easy Livin
Date Posted: September 08 2008 at 11:35
I believe the rating is affect not only by reviews of the album in question, but by reviews of other albums which contribute to the overall weighting.
 
Bear in mind too that someone else may have submitted a rating only around the same time, and that the calculations are cached, i.e. they do not always update instantly.


Posted By: Padraic
Date Posted: September 08 2008 at 11:41
Originally posted by chopper chopper wrote:

It would if the algorithm caused Roj's rating to be less than the 2.92 average.


If indeed this is the case I would claim that the algorithm is seriously flawed.


Posted By: Roj
Date Posted: September 08 2008 at 11:44
Natural Science ^ I agree entirely Clap
It would be nice if a positive review had a positive ( albeit minor I appreciate ) effect on the rating. God only knows what would have happened if I'd rated it 1 star, as opposed to 4 stars!


Posted By: MikeEnRegalia
Date Posted: September 08 2008 at 11:58
Maybe I know what's happening. If I remember correctly, the current algorithm combines the average of all the albums with the average which results from all the ratings for the album:

AVG = (R*N+r*n)/(N+n)

R = Average of all the ratings in the database (currently about 2.9 EDIT: what was the number again? Maybe M@x could clarify)
N = Average number of ratings per album
r = average rating of the album in question (arithmetical mean)
n = number of ratings for the album in question

Now, if currently r is well above 4, an additional 4 star rating might actually decrease the result.

GeekWackoLOL


-------------
https://awesomeprog.com/release-polls/pa" rel="nofollow - Release Polls

Listened to:


Posted By: Padraic
Date Posted: September 08 2008 at 12:01
Originally posted by MikeEnRegalia MikeEnRegalia wrote:

Maybe I know what's happening. If I remember correctly, the current algorithm combines the average of all the albums (it's around 3.90 if I remember correctly) with the average which results from all the ratings for the album:

AVG = (R*N+r*n)/(N+n)

R = Average of all the ratings in the database (currently about 3.9)
N = Average number of ratings per album
r = average rating of the album in question (arithmetical mean)
n = number of ratings for the album in question

Now, if currently r is well above 4, an additional 4 star rating might actually decrease the result.

GeekWackoLOL


But in this case r appears to be below 3!  Your statement is just common sense - if the album average is well above 4, a rating of exactly 4 will lower it.


Posted By: MikeEnRegalia
Date Posted: September 08 2008 at 12:02
^ I edited the post ... in retrospect I think the average is at 2.9, not 3.9. It would be useful if that value was posted somewhere on the website ... 

-------------
https://awesomeprog.com/release-polls/pa" rel="nofollow - Release Polls

Listened to:


Posted By: Roj
Date Posted: September 08 2008 at 12:19
Confused Sorry, maths never was my strong point.


Posted By: M@X
Date Posted: September 19 2008 at 01:55
Guys,

Sorry for the delay , i was out for 2 weeks.

The problem (i hope not too global) is related to a problem in older calculation of the avg. rating of the specified album in the past, data corruption. So the last review/rating posted made it recalculate correctly the new avg. rating and now giving the correct value.

I will investigate more about this..

thanks for reporting

Max


-------------
Prog On !



Print Page | Close Window

Forum Software by Web Wiz Forums® version 11.01 - http://www.webwizforums.com
Copyright ©2001-2014 Web Wiz Ltd. - http://www.webwiz.co.uk