Judging by this site, the '00s have been the weakest decade since the '50s, the era before the modern conception of the album; complex, eclectic song structure; and the popularization of foreign instruments like the zitar and modern instruments like the synthesizer. Have the last eight years been exceptionally sucky? Is this due to to a shift away from CDs to MP3s and critics as a result not coming to a consensus of what the "big" releases of any particular year are?

You are neglecting to realise that this decade has yet to be subject to decade end lists. With those submissions this decade is likely to boost considerably. Albums like Arcade Fire's Funeral, Radioheads Kid A, Sufjan Stevens Illinios and some other will likely fall in the top 50

Yeah Kevin is right. Decade lists haven't begun yet. Just as importantly, later albums haven't had the opportunity to appear on as many all-time lists.

This may be asking too much, but I think the site should take that into account. I'm sure statistics can be applied in some way to predict how albums and songs will shape up in later lists.

Keep in mind it takes a long time to fully appreciate great albums.

Nobody in the right mind in 1966 would have said "In 40 years, Pet Sounds will be considered the best album ever recorded".

Time elevates great music, and forgets average music.

I would also argue that critical opinion is a lot more divided now. There's no single type of music that's incredibly popular. Everybody goes off and listens to their own thing. Especially since the oligopoly of the major labels has been destroyed, right now there's just so much more competition for critical acclaim, the best albums aren't going to individually get as much.

Another thing that prevents newer albums from getting higher on the list is that the tie goes to the canonized.

Critics and music listeners in general already have their mental canon of greatest albums ever. So there isn't a lot of room left on that list. For a newer album to penetrate that list, the standards are higher.

Decade lists haven't begun yet. Just as importantly, later albums haven't had the opportunity to appear on as many all-time lists.

This may be asking too much, but I think the site should take that into account. I'm sure statistics can be applied in some way to predict how albums and songs will shape up in later lists.

The site does take this into account! However, BillAdama is exactly right about why the '00s is lower than the previous decades.

Maybe the decade will be considered stronger after the decade lists, not because of the statistical method, but because the critical opinion will be less divided.

VERY TRUE

Albums like: Kid A, Return to cookie Mountain, Illinios, Funeral, Is This It, In Rainbows, Elephant, Yankee Foxtrot Hotel, YS, Sound of Silver, Neon Bible.

Those titles are all going to stand out like a sore thumb when the decade lists come rolling in. We already know what the assumed greats of this decade are. Sure...We'll get some suprises, but for the most part this decade hasnt been as braod as the 70's (160 something in the top 500). This decade might include 70 if we're lucky.

But as a whole I like this decade's music. Radiohead is my favourite band, Arcade fire seem to paving way for other great canadian act, lCD Soundsystem is making some braod statments. Hip hop is starting to come around with MIA, Lil Wayne and other. I see a great next decade, I forsee a decade that revivals the 70's.

I agree with both of Bill Adama's points on this one. People generally prefer the albums and songs they already have heard 20 times to once. It takes time for popular opinion to A. listen to everything and then B. agree on which albums are best. I still think the 2000s are the worst decade so far for music (and I'm growing in this decade). There's good stuff, but there's a lack of consensus as to which stuff that is.

My expectations for this decade are (after the end of decade lists):

A. The top 5-10 or so albums will rise. These are the near-unanimous critical successes of this decade. They'll make it to the top 100 all-time.

B. The other albums to rise will be the ones that have a good ranking and were successful chart-wise. These are the Kid As, the College Dropouts, etc. The albums that were huge hits are the ones that are remembered more clearly. The ones that are really cheesy will drop like anvils, and the ones that were good, not great will rise. 17 years ago Nevermind had a solid, if not spectacular, critical rating. Today Nirvana has the #3 album and #4 song. Why? Because they were the bright spot of the early '90s mainstream. And...

C. The albums that sound most like pop music in 2010-11 will shoot up. Influence is important. Sticking with Nirvana, they wouldn't be ranked nearly as high if they didn't start a whole new genre of alternative rock. Likewise, the Pixies ratings' shot up too, because they were making good music before the kind of music they made became huge.

D. The albums that sound like "old" music will plummet. White Stripes, Amy Winehouse, this means you. Making good music is fine, but it can't be great without breaking new ground. Revivals of old music are never as good as the originals. It doesn't sound as fresh.

I don't think Elephant will fall that far, but I'm hoping Is This It does.

Every time I hear it I like it a little less.

I think Arcade Fire, Sufjan Stevens, Wilco, etc are the sort of bands that are going to last. I'm not sure about Ys. It'll probably stay relatively high, but I see it dropping some. I personally love it, but it has such major backlash a lot of lists are going to omit it completely.

I disagree about the revival stuff. Just because something came out earlier doesn't mean a new band with the same revised style can't do something to improve the sound. Look at the Vaselines. They may have been one of the first twee bands to break thanks to Kurt Cobain but Belle and Sebastian perfected the genre. Everybody seems to think My Bloody Valentine has the best album of the 90s. Where do you think they'd be if it weren't for the Cocteau Twins or the Chameleons? They aren't exactly the most acclaimed bands of the 80s.

Looking back on it. I dont think most people would consider loveless the best of the decade. OK COMPUTER usually takes that prize

I like both Loveless and Head Over Heels (Cocteau Twins) better than OK Computer. But yeah, I do see OK Computer being higher on most lists.

tim, there's a difference between the examples you gave and what I was referring to. Your examples of Belle & Sebastian and My Bloody Valentine are examples of a band being influenced by other bands that came a short time before it. What I was talking about are artists who are imitating a sound that was popular 3-4 decades ago. IMO, the White Stripes are rehashes of Led Zeppelin, i.e. and other 70's blues-rock bands. Enjoyable rehashes nonetheless, but they're not breaking any new ground.

Henrik, a question regarding your algorithm: How do you handle that some charts cover more years than others? It is obviously a lot easier to be highly ranked on a best-of-year than a best-ever list. On the other hand, it does not seem totally reasonable to weigh a best-of-decade list ten times more then the best-of-year list.

My suggestion would be to weigh lists by the square-root of the number of years they cover, i.e. a best-of-year list would have a weight of 1 whereas a best-of-decade list would have a weight of 3,16.

More to the point of this thread: A best-ever list from 1980 would have a lower weight than the same type of list from 2008. The former (basically) cover 1965-1980 (16 years), whereas the latter cover 1965-2013 (49 years). The weight of the former would then be 4 whereas the latter would have a weight of 7.

IF lists are NOT weighed this way in the current algorythm, this may be a small part of the reason why recent albums are not ranked as well as the oldies. But do be honest, I find most acclaimed albums from the last few years to be less interesting than I used to some years back. But maybe that's just me...

Robban, I don't use a point system like that. Instead, each album is matched against other albums, pair by pair, only taking into account critics lists where both albums have been eligible. The more matches won, the higher place at AM (well, that was a bit simplified, as different lists have different weights).

Example 1: In the match between "Pet Sounds" and "Kid A", only critics lists from 2001-2008 covering both 1966 and 2000 are taken into account.

Example 2: In the match between "The Marshall Mathers LP" and "Kid A", EOY lists from 2000 and lists from 2001-2008 are taken into account.

Hej igen,

Thanks for the clarification!

I probably still haven't understood it all, so another question... For the sake of the argument, if Kid A had been released in 1966 (like Pet Sounds), there are are many more eligible lists it can compete on. Would that mean that it would have had the chance to score a higher point in your system? Or does your procedure of "matches between two albums" solve this completely? (In effect, the question is: Once all games have been played, what is the procedure to aggregate them in order to prepare the list of 3000 albums?).

Sorry, once you've answered this one, I promise you that I won't bug you with any more questions in a while! :-)

Henrik, this match principle quite clearly explains why your algorithm takes a couple of hours!

If you compare, for example, 5000 albums, than you have to compute 4999*5000/2 ~ 12.5 million matches!!! That's a lot of time, especially considering that the filtering on and weighting of the lists need to be taken into account.

dr robban.

I'll answer it for him.

Its called being shit out of luck....lol

So what if Kid A wasnt made in 1966. It probably couldnt have been conceived. If Radiohead made music in the 1960's it would have been completely different.

Older albums deserve their lists because quite frankly they inspired the later music. You have to give them the extra points for doing so.

Its not to say that later music cant overcome some classics, it just has to be really good. Led Zep IV is higher then there first 2 albums. Does that make it worse?

Well, I think Henrik knows the answer to Robban's question, but I've understood that each pair of albums are being compared, so than it would not make a difference if Kid A had been released in 2000 or 1966 or any other year, it still would be compared with all other albums. But again, that's what I understood.

I quite like this system with pairwise comparisons. We could even use this for our polls here, so instead of using these sophisticated formulas ... but maybe it's too much effort for the moderators to program the algorithm every time?

Henrik, this match principle quite clearly explains why your algorithm takes a couple of hours!

Exactly, Andre! And there's nothing wrong with your answer to dr Robban either, but I'll try to explain a little more.

With all games (or matches, or whatever the proper name is) completed, I get two values for each album; the maximum possible total score and the actual total score. Then, well...a bit simplified, the final score is

final score = actual score / maximum possible score


It should be mentioned that the maximum total score is the sum of the maximum scores from every pairwise game, where

maximum score (per game) = sum of all the critics lists' weights

This means that the input on the final score is small from games with a small sum of weights (or a low number of critics lists), which is the case for all comparisons between albums from 2007 and older albums (these games do not include much more than the AMG rating).

Thanks Henrik and Andre. If the actual score is compared to the maximum possible score then the method seems to be flawless. But yeah, time-consuming!

Kevin: I never liked Led Zep. ;-)