What happened to TOP 100???

M@X wrote: Yes , there is been a upgrade to the algo , it's now based on the more adequate weighted average calculation (more info here : http://en.wikipedia.org/wiki/Weighted_average )   ;-) Any comments sor far ?

The Miracle wrote: It looks a bit weird and not as accurate as the previous one to be honest...

Seconded!

The algorithm that is used here reduces the influence of the number of ratings further compared to the former algorithm, at least that's what I get from observing the list.

The T wrote: russellk wrote: Been thinking about this change this morning. Many of us experienced prog-heads profess our disinterest in the Top 100 (doesn't stop us posting though). So who is the Top 100 for? In my opinion it serves the newer prog-lover, someone looking to build up their infant collection. That's why a weighting for popularity was very important in the old system. A newcomer to prog needs, surely, to listen to the most widely acclaimed progressive music before branching out. Yes, they should branch out, which is why the site has 'Top' lists in all the different and challenging prog genres. But listening to the top 100 should be the first aim. That 'popularity' weighting has largely gone. So what will happen now? Some of the more obscure albums lauded by a few are now in the Top 100. Newcomers to prog will purchase them. They may find the music too much for their uneducated ears, and put prog aside altogether, or at the least discard the genre from which comes the album they purchased. My point? The change is great for me: it highlights some new music to get hold of (I've not heard BACAMARTE, for example). But the Top 100 is not for me, a prog lover since the early '70s. It's for the newcomer, who will, I think, be somewhat disconcerted by some of the music they may inadvertently end up with.    This is also true... I myself did that... Followed the top-100... Now... Bacamarte? You know how easy is it to get a cd like that here in the US? PFM? (yes, with 50 USD)... We're making prog a difficult music to find for new people.... if we already agreed in that the top-100 is NOT artistically relevant, then let's make it useful for new proggsters, instead of just trying it to be as obscure as possible.

Yoy're talking like the new algo's made to include PFM in the top ten, as well as including more obscure albums in general. Its not. its obviously trying to be more accurate. If its successfull at that or not is another question.

What an album costs in the US is irrelevant. The result of that would of course mean leaving out almost every non-englishspeaking record.

Its useful learning new proggsters (especially in the UK and the US) that masterpieces of prog comes from all over the world.


Edited by Rocktopus - July 10 2007 at 18:59

From what I've seen, this new algo is much simpler.
The former considerated the note it received from all reviews (e.g: 10 reviews, 8 of 4, 1 of 3 and one of 5stars, would resul on 4.00) and also the amount of reviews.

Now, I think this new one is only judging the average note. Of course, it has a pre-requisit of a number of reviews, like 30 or 40, otherwise an album reviewed by just one person, as masterpiece, would be on top.

The T wrote: russellk wrote: Been thinking about this change this morning. Many of us experienced prog-heads profess our disinterest in the Top 100 (doesn't stop us posting though). So who is the Top 100 for? In my opinion it serves the newer prog-lover, someone looking to build up their infant collection. That's why a weighting for popularity was very important in the old system. A newcomer to prog needs, surely, to listen to the most widely acclaimed progressive music before branching out. Yes, they should branch out, which is why the site has 'Top' lists in all the different and challenging prog genres. But listening to the top 100 should be the first aim. That 'popularity' weighting has largely gone. So what will happen now? Some of the more obscure albums lauded by a few are now in the Top 100. Newcomers to prog will purchase them. They may find the music too much for their uneducated ears, and put prog aside altogether, or at the least discard the genre from which comes the album they purchased. My point? The change is great for me: it highlights some new music to get hold of (I've not heard BACAMARTE, for example). But the Top 100 is not for me, a prog lover since the early '70s. It's for the newcomer, who will, I think, be somewhat disconcerted by some of the music they may inadvertently end up with.    This is also true... I myself did that... Followed the top-100... Now... Bacamarte? You know how easy is it to get a cd like that here in the US? PFM? (yes, with 50 USD)... We're making prog a difficult music to find for new people.... if we already agreed in that the top-100 is NOT artistically relevant, then let's make it useful for new proggsters, instead of just trying it to be as obscure as possible.

I have yet to pay $50, or any price close to that for a PFM CD. Where do you shop?

OpethGuitarist wrote: Bacamarte is amazing and I'm glad they are in there (didn't notice that earlier)

Now if only I could find somewhere to BUY THEIR MUSIC!!!

Shakespeare wrote: OpethGuitarist wrote: Bacamarte is amazing and I'm glad they are in there (didn't notice that earlier) Now if only I could find somewhere to BUY THEIR MUSIC!!!

Equality 7-2521 wrote: andu wrote: Could anyone with better knowledge of math explain the new algorithm and it's characteristics? The Wikipedia article speakes Chinese to me.    If you're familar with how GPA's are calculated in college that's a weighted average.   Using wiki's notation:   Let me use the GPA as an example. X-bar would be your GPA. To find this you use two inputs, one being X. X is the grade you recieve for a given class. If you got an A, X=4. If you got a B, X=3. A C, X=2. Etc. The second input is W. W is how many credits your course is. A 4 credit course has W=4. A 3 credit course, W=3. You get the idea.   Now say you take 3 courses, Course 1 is worth 3 credits and you earn an A. Course 2 is worth 4 credits and you also earn an A. Course 3 is worth only 2 credits and you earn a B. So your values are as follows: X1=4    W1=3 X2=4    W2=4 X3=3    W3=2   so your GPA, X-bar, is calculated like so   [ (X1*W1)+(X2*W2)+(X3*W3) ] divided by W1 + W2 + W3   So your GPA will be 3.78 because you have after plugging in ( 12 + 16 + 6 ) / (3 + 4 + 2)   In the case of the albums I'm assuming X will be the rating a reviewer gives the album, and W will be the  status of the reviewer. A rating without a review being W=1, a rating with a review being W=2, and a rating with review from a collab being W=3?   Max or somebody will have to clarify exactly the inputs. I hope that clears it up.

Soul Dreamer wrote: Equality 7-2521 wrote: andu wrote: Could anyone with better knowledge of math explain the new algorithm and it's characteristics? The Wikipedia article speakes Chinese to me.    If you're familar with how GPA's are calculated in college that's a weighted average.   Using wiki's notation:   Let me use the GPA as an example. X-bar would be your GPA. To find this you use two inputs, one being X. X is the grade you recieve for a given class. If you got an A, X=4. If you got a B, X=3. A C, X=2. Etc. The second input is W. W is how many credits your course is. A 4 credit course has W=4. A 3 credit course, W=3. You get the idea.   Now say you take 3 courses, Course 1 is worth 3 credits and you earn an A. Course 2 is worth 4 credits and you also earn an A. Course 3 is worth only 2 credits and you earn a B. So your values are as follows: X1=4    W1=3 X2=4    W2=4 X3=3    W3=2   so your GPA, X-bar, is calculated like so   [ (X1*W1)+(X2*W2)+(X3*W3) ] divided by W1 + W2 + W3   So your GPA will be 3.78 because you have after plugging in ( 12 + 16 + 6 ) / (3 + 4 + 2)   In the case of the albums I'm assuming X will be the rating a reviewer gives the album, and W will be the  status of the reviewer. A rating without a review being W=1, a rating with a review being W=2, and a rating with review from a collab being W=3?   Max or somebody will have to clarify exactly the inputs. I hope that clears it up.           That's all extremely basic and straight forward, AND not different from the former algorithm! But how have THE NUMBER OF RATINGS been included here? Because I think there is the BIG difference with the former algorithm (which btw I think was better, and probably more stable for all ratings with more than 10 entries).