All Time Artists Poll: Any Takers?

[Listyguy]

As is typical for the cycle of polls, 2016 is the year of the forum's favorite artists polls. Which of course leads to the famous question: Who wants to host it? You can join a list of names that includes Moonbeam (twice I believe) and Nassim. Appologies to any who've done this poll that I forgot.

[JimmyJazz]

I haven't been on the forum in a while, but I would be really interested in hosting this. I would just need the formula and explanations for it given to me.

[Nassim]

The formula was :
points = 100 * tan (11/90) / (tan ( (10+rank)/90) )

Don't know how we got there but it worked (as long as we stay with top 100 artists at least).

As all time polls go, that's the easiest to host JimmyJazz, I ended up with 803 different artists receiving votes, that's much less than for songs or albums and there's very few cases where there's a doubt if it's the same or not.
Only issues came from cases like votes for Iggy Pop, the Stooges and Iggy and the Stooges... only situation I could not really settle was due to a vote for Curtis Mayfield and the Impressions.

[BleuPanda]

Meanwhile, the formula used in my song poll went like this:
500*(1+50)/([rank]+50)

I don't know why that was what I used, but it seemed to work. I don't know if one would work better than the other based on the number of entrants allowed in each list, but you might as well see both options (I'd personally like more than 100 artists, but whatever).

[JimmyJazz]

I personally think we should just stick with a Top 100 for our users. However, if others wish to do more than a Top 100, than feel free to chime in on this. "Artist polls" like music acts, authors, filmmakers, in my opinion, are better when pared down to just a Top 100. I personally only think a person's 100 counts for such a poll, as unlike polls of individual works, somebody's "pantheon" of artists is likely to be smaller in scale, as I can testify to myself.

In regards to Nassim's points, I think that formula works best, but again, anyone else disagrees, have their say. I also would explicitly make it clear that votes for individual performers will only count for their solo work, not for their work with groups. Groups would need to be stated otherwise.

So when do we want the deadline set for this. I am going to blast every single consistent user with a PM to alert them of the poll once I set up the thread. I really want to get a lot of people voting in this thing, as larger numbers in a poll help to create more strength behind the numbers (and also more accurately reflect our community currently).

[Listyguy]

I definitely agree. I can't fathom somebody actually having more than 100 artists they genuinely love. For me personally, I'd say I have about 50, and the next 50 on my list are at the "great but not amazing" level.

[BleuPanda]

I have made my top 100. Cutting some of these bands out was tough.

[whuntva]

I'm trying a different approach to the artists this year.

I'm going to look at my Top 500 songs and albums list, count how their appearances rank, and make a statistical average based on how well they perform.

It may take awhile, but I want to try something new.

[BleuPanda]

I have an artist file based off of one of the formulas Henrik used to use on this site, but I individually ranked the artists that placed in the top 150 to account for things such a formula would miss (i.e. LCD Soundsystem only being my 9th favorite artist with the formula simply due to their lack of albums, when they're clearly my favorite).

[jamieW]

Hopefully, the deadline won't be for a while yet. Last time, I waited till the last minute and then threw my list together based upon tracks on my I-Pod. Using this method, I managed to completely forget a few favorites that I typically listen to by streaming their music. This time, I plan on just going by an average of my song ratings (since, at the risk of sounding like a former forum member, my favorite pre-sixties artists would be penalized if I included albums), and then give "bonus points" to songs I give an 8 or higher. Needless to say, this will take some time...

[Gillingham]

I prefer not to use any statistics of albums and/or songs or anything, because for me it's more than just that. It's also their live performances, collaborations, other projects, personal attachment, ratio of good and bad songs/albums vs amount of good songs/albums, video clips, originality, and how long they've been a favorite, to name a couple of other factors.

[luney6]

I agree with Gillingham and BleuPanda.

Many artists that would otherwise have made even my top 10 will be omitted.

[Nassim]

I have an artist file based off of one of the formulas Henrik used to use on this site, but I individually ranked the artists that placed in the top 150 to account for things such a formula would miss (i.e. LCD Soundsystem only being my 9th favorite artist with the formula simply due to their lack of albums, when they're clearly my favorite).

I do that too, easier than starting from a blank slate, and less likely to allow big omissions. So I use the Henrik formula (using for albums my top 500 + a bunch of 50 albums or so that I consider tied for 501th place) and then rework it to include longevity, ethics, esthetics, live performances or even sheer sympathy (probably pushes Foo Fighters a few places up), penalties for truly terrible albums...

[JimmyJazz]

So, to restate the question, when do we want to establish a deadline for this poll?

[BleuPanda]

I think I had the songs poll open for what, 4 months? But an artist poll should be much easier to throw together, and if you do directly message people, it avoids the main concern of certain people simply missing it happen. 2 months, maybe, end of May?

[JimmyJazz]

Is 2 months too small a window for many forumers, though? Or is it just fine? Curious what others have to say about this?

[Listyguy]

Two months seems like enough time to me.

[Live in Phoenix]

What is the Henrik formula? Currently I only have about a top 10 going, so even starting a conversation with myself about the next 40 placements would be helpful...

[BleuPanda]

...I was going to copy the formula, but I realize now that I completely changed my formula.

Now I just assign a certain range of albums a certain amount of points (i.e. my top 80~ albums receive 100 points, the next batch get 75, then 60, 50, etc.), which I find is a much easier way to visualize how I'm actually comparing these acts (I do the same for songs but weigh them by .6). The Beatles having 961 points compared to Bob Dylan's 834.8 actually means a physical thing (That Bob Dylan would need the equivalent of two more albums in my top ~150 to overtake The Beatles). It's also simpler than Henrik's formula which requires exact placements, which means any movement within my list would require also adjusting the artist file. Henrik's formula works well for something updated once a year, but can get a bit out of hand. The issue with my simplified take on the idea is it doesn't differentiate between the #1 and #80 album, but when considering artists I'm more concerned with a combination of quality and quantity that it doesn't mean much.

[Maschine_Man]

I have been working on a formula that gets my Top 10 looking nice, then just assuming the rest will be okay. I put in live performance to help people like Courtney Barnett, considering I have made the effort to see her 5 times should be worth something.

Rank = (TA*377 - TAP) + (SA*20 + SB*10 + SC*5) + (L*100)

Where:

TA = Total Albums
TAP = Total of Album Positions (out of 450 or so)
SA = Songs in Top 100
SB = Songs in Top 500 (but not 100)
SC = Songs in Top 1000 (but not 500)
L = Live Performances

I was going to include Last.fm data, but I looked and it's not as accurate as I know it should be (though the current top 50 according to the formula and my last.fm top 50 share 36 names in common). Maybe I can include records/merch? Maybe a tax for those that have more than 5/6 albums?

As long as it's completely convoluted I'll be fine.


Edit: For example Julia Holter:

Points = (2*377 - (50+122) + (1*20 + 2*10 + 3*5) + (1*100)

= 737

[Henrik]

I prefer not to use any statistics of albums and/or songs or anything, because for me it's more than just that. It's also their live performances, collaborations, other projects, personal attachment, ratio of good and bad songs/albums vs amount of good songs/albums, video clips, originality, and how long they've been a favorite, to name a couple of other factors.

As much as I agree with this, I have made many ranked Spotify playlists by artists over the last couple of years and I'd like to use them for my artist ranking this time. But it wouldn't be fair to compare, say, my 20th favorite song by Portishead with my 20th favorite song by Rolling Stones. I would like to choose the songs for comparison based on the number of recorded songs by each artist. Anyone knows where I could find this information?

I would then find a reasonable formula so that if for example Rolling Stones have recorded 10 times as many songs as Portishead, I would compare my 15th favorite song by Portishead with my 25th favorite song by Rolling Stones. Or something like that.

However, I would probably also need to tweak the formula for guys like Ennio Morricone who has probably made thousands of recordings. I guess what I really want is a number of "high to medium profile" songs by each artist. Any help would be really appreciated!

[BleuPanda]

For me, I only ever pay attention to the good an artist puts out. Otherwise Nirvana and Joy Division get an advantage due to untimely deaths, and The Rolling Stones are hurt simply due to surviving.

[Gillingham]

For me, I only ever pay attention to the good an artist puts out. Otherwise Nirvana and Joy Division get an advantage due to untimely deaths, and The Rolling Stones are hurt simply due to surviving.

That would not be a problem for me, since Nirvana and Joy Division don't have a high amount of (good) work, whereas the Stones have.