Progarchives.com has always (since 2002) relied on banners ads to cover web hosting fees and all.
Please consider supporting us by giving monthly PayPal donations and help keep PA fast-loading and ad-free forever.
/PAlogo_v2.gif "Progarchives.com Homepage") |
|
Post Reply 
|
Page <1 23456 15> |
| Author | ||||||||||||||||||
tuxon 
Forum Senior Member 
Joined: September 21 2004 Location: plugged-in Status: Offline Points: 5502 |
 Posted: July 10 2007 at 21:31 |
|||||||||||||||||
|
what is currently calculated? a weighted mean only has value if a non-disputable value can be derived from that mean, but all that happens is a calculation of a mean and that's given a non-specific value based on argumental weighing based on nothing, though maybe the result looks acceptable, it is neither a meisuring of quality nor of popularity, it just happens to look good becuse every weighted mean will benefit a large number of ratings above lower numbers of data.
basically it hasn't improved anything we had in the past, and made it more difficult to read the meaning one could give to such list.
if you wan to use such logarhytm, please calculate the mean - 2,5 that would help a bit, without really comming to any worthfull conclusions.
previous algorhythm wasn't perfect, but could be understood, currently it's just a difficult calculation signifying nothing.
|
||||||||||||||||||
|
I'm always almost unlucky _ _ _ _ _ _ _ _ Id5ZcnjXSZaSMFMC Id5LM2q2jfqz3YxT
|
||||||||||||||||||
 |
||||||||||||||||||
|
StyLaZyn
Forum Senior Member
Joined: November 22 2005 Location: United States Status: Offline Points: 4079 |
Posted: July 10 2007 at 21:36 |
|||||||||||||||||
|
I think it looks more realistic. Good job!
|
||||||||||||||||||
|
||||||||||||||||||
|
||||||||||||||||||
|
Soul Dreamer
Forum Senior Member
Joined: October 17 2005 Location: Netherlands Status: Offline Points: 997 |
Posted: July 10 2007 at 21:46 |
|||||||||||||||||
|
^^^ What is realistic? That the records you like are more in front? Can be but to my humble opinion the popularity of a record (=the number of entries) should be incorporated as well, and I think now that's underrated.
|
||||||||||||||||||
|
To be the one who seeks so I may find .. (Metallica)
|
||||||||||||||||||
|
||||||||||||||||||
|
explodingjosh
Forum Senior Member
Joined: February 10 2007 Location: United States Status: Offline Points: 507 |
Posted: July 10 2007 at 22:03 |
|||||||||||||||||
|
I think it sucks balls.
 Why? Well it doesn't even make sense really, because I'm looking at positions number 43, 44, 45 right now ( I will not even mention which albums they are because I want to be completely objective in my argument ) and they all have the same rating. But even though 45 has a greater number of ratings than 44, and 44 greater than 43, their order does not reflect that. If they have the same star rating, the one with more ratings should be higher. That is not the case with 43, 44, 45, 51 & 53, 60 & 61and others. I argue this purely from a statistical standpoint; I could care less which actual albums are in what position. ...as long as pink floyd is #1  |
||||||||||||||||||
|
||||||||||||||||||
|
Soul Dreamer
Forum Senior Member
Joined: October 17 2005 Location: Netherlands Status: Offline Points: 997 |
Posted: July 10 2007 at 22:14 |
|||||||||||||||||
|
I don't want to impose on this thread, but I think the former algorithm was near to perfection (thanks Mike!). The ONLY place it let down, was with the albums with low number of entries (1-10). There you could get the strange phenomenon that an album with 1 rating of 5* would be higher than say another album with 4 ratings having 4,3 as average, which is not as it should be. That however, can be solved very easilly by applying a kind of Michaelis-Menten or (for the chemists Langmuir) isotherm approach, which would take away this strange situation.
Now I feel that the algorithm used is to calculate the weighted average, and then deduct something from the score for the number of entries (more entries = less deduction). I just think the more "analog" approach as divised by Mike is better.
|
||||||||||||||||||
|
To be the one who seeks so I may find .. (Metallica)
|
||||||||||||||||||
|
||||||||||||||||||
|
Evandro Martini
Forum Senior Member
Joined: February 08 2006 Status: Offline Points: 183 |
Posted: July 10 2007 at 23:34 |
|||||||||||||||||
|
I agree with everything you've said!
|
||||||||||||||||||
|
"You’ll never make any money playing music that people can’t sing.” Keith Emerson's father
|
||||||||||||||||||
|
||||||||||||||||||
|
Dim
Prog Reviewer 
Joined: April 17 2007 Location: Austin TX Status: Offline Points: 6890 |
Posted: July 11 2007 at 00:35 |
|||||||||||||||||
Go join the punk rock archives!
|
||||||||||||||||||
|
||||||||||||||||||
|
||||||||||||||||||
|
Sckxyss
Forum Senior Member
Joined: May 05 2007 Location: Canada Status: Offline Points: 1319 |
Posted: July 11 2007 at 00:52 |
|||||||||||||||||
|
Perhaps it works as Equality described, but with a bunch of default hidden ratings given to albums, to make the first bunch of ratings have less of an effect. Then, the only way for an album to get into the 4s would be for it to have a significant number of high ratings. That would explain how most of the albums with only a few ratings all have averages of around 3.5
That's the best I can come up with, from what I've seen  EDIT: Scratch that, the quartermass album someone mentioned earlier with only one review of 1 star has an average that's HIGHER than albums with only one 4 star rating. I'm completely stumped. Edited by Sckxyss - July 11 2007 at 00:58 |
||||||||||||||||||
|
||||||||||||||||||
|
King Lerxt
Forum Newbie 
Joined: August 15 2006 Status: Offline Points: 28 |
Posted: July 11 2007 at 01:10 |
|||||||||||||||||
|
I really don't like it... Sorry!
What happened to Snakes and Arrows listed on 2007 top 100??? |
||||||||||||||||||
|
||||||||||||||||||
|
MikeEnRegalia
Special Collaborator 
Honorary Collaborator Joined: April 22 2005 Location: Sweden Status: Offline Points: 21206 |
Posted: July 11 2007 at 02:59 |
|||||||||||||||||
Wow ... I have a degree in computational engineering (which includes a great deal of mathematics) and I can't understand your reasoning at all. Why should a weighted mean "benefit" from a large number of ratings? It doesn't. The resulting mean is an average, which by definition doesn't depend on the number of values. And the weights aren't based on "nothing" - they're based on reviewer/review rank (special collab, review, rating without text). Essentially what has happened is that the number of reviews does not have any significance anymore (since it is not used in the computation), and the resulting averages are now compared linearly (in the previous formula avg^3 was used). |
||||||||||||||||||
|
||||||||||||||||||
|
richardh
Prog Reviewer
Joined: February 18 2004 Location: United Kingdom Status: Offline Points: 28270 |
Posted: July 11 2007 at 03:23 |
|||||||||||||||||
Brain Salad Surgery not even top 100   Edited by richardh - July 11 2007 at 03:23 |
||||||||||||||||||
|
||||||||||||||||||
|
richardh
Prog Reviewer
Joined: February 18 2004 Location: United Kingdom Status: Offline Points: 28270 |
Posted: July 11 2007 at 03:24 |
|||||||||||||||||
|
Well at least now we can get away from the pretence that this top 100 means anything
|
||||||||||||||||||
|
||||||||||||||||||
|
Logan
Forum & Site Admin Group
Site Admin Joined: April 05 2006 Location: Vancouver, BC Status: Offline Points: 36152 |
Posted: July 11 2007 at 03:25 |
|||||||||||||||||
|
Mike: Math is not my strong suit (and it's late and I'm too tired to try to figure it out -- plus this site is not agreeing with my computer so pages time out), so I wonder why do all the unrated albums in the top 4000 now have a value of 3.86 (by my fatigued checking)? I have a theory, but...
I don't find that useful. Taken from the top 4000 most popular albums (CLICK): Every non-rated album in the top 4000 is given a rating of 3.86 and they all fall between 1224 and 2533 (they are surrounded by reviewed albums). Between 2882 and 3286 all the albums have 2 ratings and are given 3.83. Between 2881 and 3153, the first albums only rated once are found, and are given 3.84. And at 3379 to 4000 are all the other albums given one star - at 3.83. All of the albums only rated three times are given a score of 3.85, 3.84, or 3.83. The album at 4000 with one review/rating of 4 by Sean Trane which gets a list rating of 3.83. I see looking at the Kevin Ayers page that the album ratings do vary slightly between the albums only rated once... from 3.79 to 3.83. Actually, I know notice for less variation over all in the rating of albums from band pages. As has been mentioned, low-rated ones are bumped up and high-rated ones go down. King Lerxt: Rush is now at 162. Asia's one review was at 2 stars, and the other is a rating without review. Xang has a 3 and a 5 (from collabs), and a 3 from another reviewer, plus one rating without review. Incidentally, Asia has three other reviews with a 2 star review each that score 3.70, and others with but one three star review each score 3.82.
Edited by Logan - July 11 2007 at 03:27 |
||||||||||||||||||
|
||||||||||||||||||
|
russellk
Prog Reviewer
Joined: February 28 2005 Location: New Zealand Status: Offline Points: 782 |
Posted: July 11 2007 at 03:29 |
|||||||||||||||||
|
Press the reset button, PA bosses; this obviously hasn't worked.
|
||||||||||||||||||
|
||||||||||||||||||
|
fuxi
Prog Reviewer
Joined: March 08 2006 Location: United Kingdom Status: Offline Points: 2459 |
Posted: July 11 2007 at 03:30 |
|||||||||||||||||
Me neither! I couldn't tell you if this way of measuring is better or not, but it still looks like a an incredibly thrilling Top-100 to me. Great to be part of such an amazing site! 
P.S. I do agree about leaving out proto-prog and prog-related bands, though... Edited by fuxi - July 11 2007 at 03:33 |
||||||||||||||||||
|
||||||||||||||||||
|
aprusso
Forum Senior Member
Joined: June 16 2005 Status: Offline Points: 312 |
Posted: July 11 2007 at 03:35 |
|||||||||||||||||
|
So what's the new algorithm? This top-100 seems absolutely incoherent. the one before was so much more reflecting values at stake
|
||||||||||||||||||
|
||||||||||||||||||
|
andu
Forum Senior Member
Joined: September 27 2006 Location: Romania Status: Offline Points: 3089 |
Posted: July 11 2007 at 04:00 |
|||||||||||||||||
Thanks! We don't have a GPA (and I don't know what the acronym stands for), in my country we just sum up the number of credits multiplied with the grades to count the whole year's "weight". Let's now wait to find out what X and W are... |
||||||||||||||||||
|
||||||||||||||||||
|
Rocktopus
Forum Senior Member
Joined: March 02 2006 Location: Norway Status: Offline Points: 4202 |
Posted: July 11 2007 at 04:29 |
|||||||||||||||||
There's obviously a lot of stuff that's wrong here. Could the new algorithm be perfected/fixed somehow? If not, I got to agree that we should go back to using the previous one. Not because of what albums are and are not in the new TOP 100. Some reviews not being counted at all is an insult. And the rest of what Logan's pointing out is completely absurd. |
||||||||||||||||||
|
Over land and under ashes
In the sunlight, see - it flashes Find a fly and eat his eye But don't believe in me Don't believe in me Don't believe in me |
||||||||||||||||||
|
||||||||||||||||||
|
Andrea Cortese
Special Collaborator
Honorary Collaborator Joined: September 05 2005 Status: Offline Points: 4411 |
Posted: July 11 2007 at 04:30 |
|||||||||||||||||
|
I don't understand why the most part of the albums of the ISP genre has been downrated. Cherry Five, LuxAde, Nostos and many other are between 3 and 4!!
|
||||||||||||||||||
|
||||||||||||||||||
|
Andrea Cortese
Special Collaborator
Honorary Collaborator Joined: September 05 2005 Status: Offline Points: 4411 |
Posted: July 11 2007 at 04:34 |
|||||||||||||||||
In my opinion that's a correct move. It's important not to have in the main page only the most popular albums. This site's objective is to open the door to the huge amount of prog bands and artists that risk to remain outside the official "best of" poll.
The only remark, as I told before, is that I don't understand why so many albums has been downrated. Hybla Act I by Randone (2005), for example. The 73% of people has rated it with 5 stars. The finaly rating is below 4. Before the site's changings it was 4,30.
|
||||||||||||||||||
|
||||||||||||||||||
|
Post Reply
|
Page <1 23456 15> |
| Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |
What happened to TOP 100???
Topic Options
The T wrote:
. And indeed if this thread wasn't here i wouldn't have noticed for a month.