UPDATE: The poll has been closed. Click here to see the results:

http://www.cs.cornell.edu...thm=runoff

Hi again:

Condorcet Album Poll: Crystal Ball

This is the next installment in a series of polls ranking the songs on each album officially released by Prince or O+>. The goal of this project is to determine which tracks should be included in a final poll to determine the overall top 25 Prince songs in the opinion of his fans.

For those of you who haven’t been following the polls, these are different than the usual “best song” threads in that they combine the preferences of all the voters into an overall list.

I am going through these in the same order that they were ranked in the earlier poll of the best Prince albums. Now's the time of Crystal Ball (the official 1998 release, not the 1986 unreleased album of the same name). 30 songs to rank.

Please don’t discuss the preliminary results or campaign for your favorite song(s) until the poll is closed. I plan to close the poll on Tuesday, but I may close it early if the preliminary results are discussed on this site. This is to make sure the poll is as accurate as possible.

Click here to vote in the current poll:

http://www.cs.cornell.edu...5a1394a784

For the previous polls, click here:

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu/w8/~andru/cgi-perl/civs/vote.pl?id=E_869001eb46cf806c&akey=5651316958e6bd23

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...268d5a19ee

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...e60c4099c9

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

http://www.cs.cornell.edu...thm=runoff

Also, if you haven't yet, please nominate your favorite Prince songs that were not released on Prince's albums or singles here:

prince.org/msg/7/364193

Thanks!

[Edited 8/23/11 7:28am]

What's the ultimate goal of these surveys? Does it relate in any way to future Prince releases?

Fauxie said:

What's the ultimate goal of these surveys? Does it relate in any way to future Prince releases?

The results of all the polls will determine which tracks should be included in a final poll to decide the overall top 25 Prince songs in the opinion of his fans. So when all the album polls are completed, and also a poll of Prince's songs that were not released in Prince's albums or singles, there'll be a final poll to determine a list with the 25 Prince songs most loved by his fans.

Javi said:

The results of all the polls will determine which tracks should be included in a final poll to decide the overall top 25 Prince songs in the opinion of his fans. So when all the album polls are completed, and also a poll of Prince's songs that were not released in Prince's albums or singles, there'll be a final poll to determine a list with the 25 Prince songs most loved by his fans.

Thanks for the explanation. I enjoy them and the results are interesting. I just never knew if there was an ultimate goal that was to be accomplished by taking the surveys. Thanks again for the clarification.

Yeah, I wasn't poo poo-ing the surveys. Just wondering.

Got me thinking about Crystal Ball 2 or something like that. A Prince-endorsed Greatest Hits package to come.

Of course, as Imago would say, Prince hates his fans, and we love him for it, so that was silly on my part.

.

[Edited 8/21/11 7:44am]

Aristotle said:

Javi said:

The results of all the polls will determine which tracks should be included in a final poll to decide the overall top 25 Prince songs in the opinion of his fans. So when all the album polls are completed, and also a poll of Prince's songs that were not released in Prince's albums or singles, there'll be a final poll to determine a list with the 25 Prince songs most loved by his fans.

Thanks for the explanation. I enjoy them and the results are interesting. I just never knew if there was an ultimate goal that was to be accomplished by taking the surveys. Thanks again for the clarification.

Thank you for taking part.

There's the described ultimate goal, but I agree that the polls themselves are very interesting, because, as you must rank all the songs in each album, and not only say which song is your favourite, it's a good chance to think again about the records, or even listen to them. And it's always interesting to see what the other orgers think, and to compare your views with the final results.

How are you going to weigh the top tracks from each album against each other? Surely the #5 track on Purple Rain would still rate higher than the #1 track on Chaos and Disorder.

UltraLash said:

How are you going to weigh the top tracks from each album against each other? Surely the #5 track on Purple Rain would still rate higher than the #1 track on Chaos and Disorder.

Actually, although I'm conducting them for the time being, the idea of the polls came from another orger, electricberet, so he could explain this better.

But you're right with your comment. In the beginning there was a poll to rank all the albums, so in the final poll there will certainly be more songs of the album that ranked higher (Sign O' The Times) than of the other albums.

Anyway, I think the songs that win in their corresponding album polls will be considered for the final poll, even the one in the album that ranked the lowest (Rave Un2 The Joy Fantastic).

You can see the results of the albums poll here:

http://www.cs.cornell.edu...thm=runoff

Also, there was a singles poll which can serve to measure the results of each album poll. With an example: I Would Die 4 U ranked 17 in the singles poll, but in the Purple Rain album poll there were three tracks which weren't singles that ranked higher: The Beautiful Ones, Computer Blue and Darling Nikki. Since I Would Die 4 U's ranking in the singles poll was relatively high, these three songs must be in the final poll too, as they are more loved by the orgers according to the Purple Rain poll.

These are the results of the singles poll:

http://www.cs.cornell.edu...thm=runoff

And these the results of the Purple Rain poll:

http://www.cs.cornell.edu...thm=runoff

I hope my explanation has clarified, at least in part, your doubt.

UltraLash said:

How are you going to weigh the top tracks from each album against each other? Surely the #5 track on Purple Rain would still rate higher than the #1 track on Chaos and Disorder.

I have been thinking about that ever since I came up with the idea for these polls, but I haven't come to a resolution. It seems fair that more songs should be included in the final poll from the albums that ranked higher in the album poll (thus SOTT should have more entrants than Rave). The songs that ranked well in the A-sides and B-sides poll should also be included, along with the winners of the unreleased/associated-artist poll that has yet to be conducted. But if the list of songs in the final poll gets too long, it will be hard to rank them all and fewer people will vote. It might be necessary to have one or more runoff polls before we do the final one. Some of the early polls didn't have as many voters as the later ones, and so the results may not reflect the opinion of the org as a whole.

When all the preliminary polls have been finished, maybe we should start a thread to discuss the methodology for the final poll. We could even do a poll to determine the best method of doing the final poll! After all, we have nothing better to do, right?

electricberet said:

UltraLash said:

How are you going to weigh the top tracks from each album against each other? Surely the #5 track on Purple Rain would still rate higher than the #1 track on Chaos and Disorder.

I have been thinking about that ever since I came up with the idea for these polls, but I haven't come to a resolution. It seems fair that more songs should be included in the final poll from the albums that ranked higher in the album poll (thus SOTT should have more entrants than Rave). The songs that ranked well in the A-sides and B-sides poll should also be included, along with the winners of the unreleased/associated-artist poll that has yet to be conducted. But if the list of songs in the final poll gets too long, it will be hard to rank them all and fewer people will vote. It might be necessary to have one or more runoff polls before we do the final one. Some of the early polls didn't have as many voters as the later ones, and so the results may not reflect the opinion of the org as a whole.

When all the preliminary polls have been finished, maybe we should start a thread to discuss the methodology for the final poll. We could even do a poll to determine the best method of doing the final poll! After all, we have nothing better to do, right?

I always suspected that there's nothing Prince fans love more than a good poll.

* just a joke, no flames please

electricberet said:

UltraLash said:

How are you going to weigh the top tracks from each album against each other? Surely the #5 track on Purple Rain would still rate higher than the #1 track on Chaos and Disorder.

I have been thinking about that ever since I came up with the idea for these polls, but I haven't come to a resolution. It seems fair that more songs should be included in the final poll from the albums that ranked higher in the album poll (thus SOTT should have more entrants than Rave). The songs that ranked well in the A-sides and B-sides poll should also be included, along with the winners of the unreleased/associated-artist poll that has yet to be conducted. But if the list of songs in the final poll gets too long, it will be hard to rank them all and fewer people will vote. It might be necessary to have one or more runoff polls before we do the final one. Some of the early polls didn't have as many voters as the later ones, and so the results may not reflect the opinion of the org as a whole.

When all the preliminary polls have been finished, maybe we should start a thread to discuss the methodology for the final poll. We could even do a poll to determine the best method of doing the final poll! After all, we have nothing better to do, right?

electricberet said:

UltraLash said:

How are you going to weigh the top tracks from each album against each other? Surely the #5 track on Purple Rain would still rate higher than the #1 track on Chaos and Disorder.

I have been thinking about that ever since I came up with the idea for these polls, but I haven't come to a resolution. It seems fair that more songs should be included in the final poll from the albums that ranked higher in the album poll (thus SOTT should have more entrants than Rave). The songs that ranked well in the A-sides and B-sides poll should also be included, along with the winners of the unreleased/associated-artist poll that has yet to be conducted. But if the list of songs in the final poll gets too long, it will be hard to rank them all and fewer people will vote. It might be necessary to have one or more runoff polls before we do the final one. Some of the early polls didn't have as many voters as the later ones, and so the results may not reflect the opinion of the org as a whole.

When all the preliminary polls have been finished, maybe we should start a thread to discuss the methodology for the final poll. We could even do a poll to determine the best method of doing the final poll! After all, we have nothing better to do, right?

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

Up to 31 votes! Keep going!

Javi said:

electricberet said:

I have been thinking about that ever since I came up with the idea for these polls, but I haven't come to a resolution. It seems fair that more songs should be included in the final poll from the albums that ranked higher in the album poll (thus SOTT should have more entrants than Rave). The songs that ranked well in the A-sides and B-sides poll should also be included, along with the winners of the unreleased/associated-artist poll that has yet to be conducted. But if the list of songs in the final poll gets too long, it will be hard to rank them all and fewer people will vote. It might be necessary to have one or more runoff polls before we do the final one. Some of the early polls didn't have as many voters as the later ones, and so the results may not reflect the opinion of the org as a whole.

When all the preliminary polls have been finished, maybe we should start a thread to discuss the methodology for the final poll. We could even do a poll to determine the best method of doing the final poll! After all, we have nothing better to do, right?

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

I agree about repeating the SOTT poll. It's the one poll I missed as well. As for a method, how about this...

If we're looking for an eventual Top 25, then create a final list composed of:

1. The top 25 singles from the Singles Poll

2. Top 15 B-Sides from the B-Sides Poll

3. Top 10 Non-album/Non-b-side/Fan Releases (Was there a Poll for this?)

4. The top songs from all 30 albums that did not rank in the top 25 singles or b-sides. And you'd choose those top songs like this: Divide all 30 albums into 5 groups (of 6 albums each). The top 5 albums will get their top 5 songs added to the list (minus the singles that have already been chosen). -- Example: For the Purple Rain album, that means that Doves, Purple Rain, Let's Go Crazy, and I Would Die 4 U make it in because they're already in the top 25 singles, and we add The Beautiful Ones and Computer Blue because they're in the top 5 on the album. -- The next group of 5 albums will get their top 4 songs added to the list (minus the singles that have already been chosen). The next group gets three songs. The next group gets two. And the final group only gets one. Or, if you want to be a little more forgiving to the lower-ranking albums, the final two groups of albums could get two selections instead of two and one.

You're bound to end up with a list that's about 110 to 115 songs long to narrow down to 25.

Damn, I put a lot of thought into that.

[Edited 8/22/11 9:44am]

UltraLash said:

Javi said:

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

I agree about repeating the SOTT poll. It's the one poll I missed as well. As for a method, how about this...

If we're looking for an eventual Top 25, then create a final list composed of:

1. The top 25 singles from the Singles Poll

2. Top 25 B-Sides from the B-Sides Poll

3. The top songs from all 30 albums that did not rank in the top 25 singles or b-sides. And you'd choose those top songs like this: Divide all 30 albums into 5 groups (of 6 albums each). The top 5 albums will get their top 5 songs added to the list (minus the singles that have already been chosen). -- Example: For the Purple Rain album, that means that Doves, Purple Rain, Let's Go Crazy, and I Would Die 4 U make it in because they're already in the top 25 singles, and we add The Beautiful Ones and Computer Blue because they're in the top 5 on the album. -- The next group of 5 albums will get their top 4 songs added to the list (minus the singles that have already been chosen). The next group gets three songs. The next group gets two. And the final group only gets one. Or, if you want to be a little more forgiving to the lower-ranking albums, the final two groups of albums could get two selections instead of two and one.

You're bound to end up with a list that's about 100 to 115 songs long to narrow down to 25.

Damn, I put a lot of thought into that.

Thanks a lot for the input! This certainly deserves to be thought about. Don't forget the poll of the Prince songs that weren't published in his albums or singles, which should contribute with some outtakes and related artists' songs.

Javi said:

UltraLash said:

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

I agree about repeating the SOTT poll. It's the one poll I missed as well. As for a method, how about this...

If we're looking for an eventual Top 25, then create a final list composed of:

1. The top 25 singles from the Singles Poll

2. Top 25 B-Sides from the B-Sides Poll

3. The top songs from all 30 albums that did not rank in the top 25 singles or b-sides. And you'd choose those top songs like this: Divide all 30 albums into 5 groups (of 6 albums each). The top 5 albums will get their top 5 songs added to the list (minus the singles that have already been chosen). -- Example: For the Purple Rain album, that means that Doves, Purple Rain, Let's Go Crazy, and I Would Die 4 U make it in because they're already in the top 25 singles, and we add The Beautiful Ones and Computer Blue because they're in the top 5 on the album. -- The next group of 5 albums will get their top 4 songs added to the list (minus the singles that have already been chosen). The next group gets three songs. The next group gets two. And the final group only gets one. Or, if you want to be a little more forgiving to the lower-ranking albums, the final two groups of albums could get two selections instead of two and one.

You're bound to end up with a list that's about 100 to 115 songs long to narrow down to 25.

Damn, I put a lot of thought into that.

Thanks a lot for the input! This certainly deserves to be thought about. Don't forget the poll of the Prince songs that weren't published in his albums or singles, which should contribute with some outtakes and related artists' songs.

Yes, I JUST went back and updated my formula with that, but you beat me to it!

UltraLash said:

Javi said:

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

I agree about repeating the SOTT poll. It's the one poll I missed as well. As for a method, how about this...

If we're looking for an eventual Top 25, then create a final list composed of:

1. The top 25 singles from the Singles Poll

2. Top 15 B-Sides from the B-Sides Poll

3. Top 10 Non-album/Non-b-side/Fan Releases (Was there a Poll for this?)

4. The top songs from all 30 albums that did not rank in the top 25 singles or b-sides. And you'd choose those top songs like this: Divide all 30 albums into 5 groups (of 6 albums each). The top 5 albums will get their top 5 songs added to the list (minus the singles that have already been chosen). -- Example: For the Purple Rain album, that means that Doves, Purple Rain, Let's Go Crazy, and I Would Die 4 U make it in because they're already in the top 25 singles, and we add The Beautiful Ones and Computer Blue because they're in the top 5 on the album. -- The next group of 5 albums will get their top 4 songs added to the list (minus the singles that have already been chosen). The next group gets three songs. The next group gets two. And the final group only gets one. Or, if you want to be a little more forgiving to the lower-ranking albums, the final two groups of albums could get two selections instead of two and one.

You're bound to end up with a list that's about 110 to 115 songs long to narrow down to 25.

Damn, I put a lot of thought into that.

[Edited 8/22/11 9:44am]

This poll will take place after the album polls are completed. By the way, you can nominate songs for this poll in this thread:

prince.org/msg/7/364193

UltraLash said:

Javi said:

I agree with the idea of one or more runoff polls. So could many songs be ranked and, at the same time, the final poll would have less songs. In my opinion, this project deserves to have a big list as a result, maybe 50 songs? That would allow space for surprises, songs from periods that in general may not be much appreciated by the fans... A list with 25 songs may include only "expected" or very well known songs, and the possibilities of a rich discussion of the results could decrease.

Also, as the number of voters have been increasing (I was gladly surprised of how many people voted in the Emancipation poll), since there were 36 songs, and of an album that didn't rank particularly high), I thinks we may expect a high number of votes in the final poll, even if more songs are included.

One last thing: I'd suggest that the Sign O' The Times poll is repeated. It's the one in which less people voted, as it had to be interrupted due to people discussing the results before the poll was closed. In my opinion, it would be a shame that the poll of the album that ranked the highest has the lowest number of voters.

I agree about repeating the SOTT poll. It's the one poll I missed as well. As for a method, how about this...

If we're looking for an eventual Top 25, then create a final list composed of:

1. The top 25 singles from the Singles Poll

2. Top 15 B-Sides from the B-Sides Poll

3. Top 10 Non-album/Non-b-side/Fan Releases (Was there a Poll for this?)

4. The top songs from all 30 albums that did not rank in the top 25 singles or b-sides. And you'd choose those top songs like this: Divide all 30 albums into 5 groups (of 6 albums each). The top 5 albums will get their top 5 songs added to the list (minus the singles that have already been chosen). -- Example: For the Purple Rain album, that means that Doves, Purple Rain, Let's Go Crazy, and I Would Die 4 U make it in because they're already in the top 25 singles, and we add The Beautiful Ones and Computer Blue because they're in the top 5 on the album. -- The next group of 5 albums will get their top 4 songs added to the list (minus the singles that have already been chosen). The next group gets three songs. The next group gets two. And the final group only gets one. Or, if you want to be a little more forgiving to the lower-ranking albums, the final two groups of albums could get two selections instead of two and one.

You're bound to end up with a list that's about 110 to 115 songs long to narrow down to 25.

Damn, I put a lot of thought into that.

[Edited 8/22/11 9:44am]

I think this could be a good framework. We might want to proceed in stages, however. The first poll could include, say, songs 20-35 from the A-sides poll and the songs that beat those songs in the other polls but did not beat the top 20 songs in the A-sides poll. You could also include the top song from each album that had no singles in the top 20 of the A-side poll. Based on that poll, you should be able to eliminate some songs and albums from consideration, and then the top 25 would advance to the next round. How many rounds prior to the final round would depend on how many songs are eliminated by the first poll. For example, if the top song on The Black Album (which had no singles) doesn't make the top 25 in the first-round poll, there's no need to poll other songs on that album in a subsequent poll. (I say this as a huge fan of The Black Album--I think it's quite possible that more than one track from that album could make the final top 25, but we'll see.)

I'm just throwing these numbers out without looking at all the polls. But the idea is to choose a group of songs that came close to making the top 25 by some measure and then pick the best from that group to add to the final poll. Sort of like a sports tournament with multiple elimination rounds. If the final poll includes more than 100 songs, we aren't going to get a lot of voters because it will take too long to complete, especially for orgers who aren't intimately familiar with all the albums. Imagine trying to compare 100 songs in one poll from different phases of Prince's long career, some of them unreleased--not an easy task! You want each poll to include a manageable number of songs. On the other hand, we have to keep in mind that some of the polls didn't get a lot of votes and may not be that accurate. The A-sides poll, for example, doesn't always match the ordering of songs in the individual album polls with respect to A-sides from each album.

I don't see any harm in repeating the SOTT poll. We could include the top songs from the original poll along with the top songs from the revote, so that people who voted either time would have their voice heard. When I did that poll, people weren't familiar with the rules so I had to end it early before it turned into a fight between those who like "The Cross" and those who don't.

One thing I would say for sure is that the final poll should only be conducted once, it should be advertised well in advance (maybe made a sticky), and it should be kept open for longer than the other polls if possible. Once that final poll has been conducted it will influence people's thinking one way or the other, so we need to get it right the first time.

electricberet said:

UltraLash said:

I agree about repeating the SOTT poll. It's the one poll I missed as well. As for a method, how about this...

If we're looking for an eventual Top 25, then create a final list composed of:

1. The top 25 singles from the Singles Poll

2. Top 15 B-Sides from the B-Sides Poll

3. Top 10 Non-album/Non-b-side/Fan Releases (Was there a Poll for this?)

4. The top songs from all 30 albums that did not rank in the top 25 singles or b-sides. And you'd choose those top songs like this: Divide all 30 albums into 5 groups (of 6 albums each). The top 5 albums will get their top 5 songs added to the list (minus the singles that have already been chosen). -- Example: For the Purple Rain album, that means that Doves, Purple Rain, Let's Go Crazy, and I Would Die 4 U make it in because they're already in the top 25 singles, and we add The Beautiful Ones and Computer Blue because they're in the top 5 on the album. -- The next group of 5 albums will get their top 4 songs added to the list (minus the singles that have already been chosen). The next group gets three songs. The next group gets two. And the final group only gets one. Or, if you want to be a little more forgiving to the lower-ranking albums, the final two groups of albums could get two selections instead of two and one.

You're bound to end up with a list that's about 110 to 115 songs long to narrow down to 25.

Damn, I put a lot of thought into that.

[Edited 8/22/11 9:44am]

I think this could be a good framework. We might want to proceed in stages, however. The first poll could include, say, songs 20-35 from the A-sides poll and the songs that beat those songs in the other polls but did not beat the top 20 songs in the A-sides poll. You could also include the top song from each album that had no singles in the top 20 of the A-side poll. Based on that poll, you should be able to eliminate some songs and albums from consideration, and then the top 25 would advance to the next round. How many rounds prior to the final round would depend on how many songs are eliminated by the first poll. For example, if the top song on The Black Album (which had no singles) doesn't make the top 25 in the first-round poll, there's no need to poll other songs on that album in a subsequent poll. (I say this as a huge fan of The Black Album--I think it's quite possible that more than one track from that album could make the final top 25, but we'll see.)

I'm just throwing these numbers out without looking at all the polls. But the idea is to choose a group of songs that came close to making the top 25 by some measure and then pick the best from that group to add to the final poll. Sort of like a sports tournament with multiple elimination rounds. If the final poll includes more than 100 songs, we aren't going to get a lot of voters because it will take too long to complete, especially for orgers who aren't intimately familiar with all the albums. Imagine trying to compare 100 songs in one poll from different phases of Prince's long career, some of them unreleased--not an easy task! You want each poll to include a manageable number of songs. On the other hand, we have to keep in mind that some of the polls didn't get a lot of votes and may not be that accurate. The A-sides poll, for example, doesn't always match the ordering of songs in the individual album polls with respect to A-sides from each album.

I don't see any harm in repeating the SOTT poll. We could include the top songs from the original poll along with the top songs from the revote, so that people who voted either time would have their voice heard. When I did that poll, people weren't familiar with the rules so I had to end it early before it turned into a fight between those who like "The Cross" and those who don't.

One thing I would say for sure is that the final poll should only be conducted once, it should be advertised well in advance (maybe made a sticky), and it should be kept open for longer than the other polls if possible. Once that final poll has been conducted it will influence people's thinking one way or the other, so we need to get it right the first time.

I totally agree with this. I had also thought about the sticky and the longer duration. As people have become familiar with the norms, I don't think it'll cause problems to have the poll open for, say, four days.

This methodology is quite interesting also in that it allows to rank a great number of songs while, at the same time, having an affordable final poll. I mean, the songs that are eliminated in the, say, third round (just a guess) will turn out ranked and could be added to the results of the final 25 songs-poll. So could we have a big list and, at the same time, polls with not too many songs, so that the number of voters is high.

Javi said:

electricberet said:

I think this could be a good framework. We might want to proceed in stages, however. The first poll could include, say, songs 20-35 from the A-sides poll and the songs that beat those songs in the other polls but did not beat the top 20 songs in the A-sides poll. You could also include the top song from each album that had no singles in the top 20 of the A-side poll. Based on that poll, you should be able to eliminate some songs and albums from consideration, and then the top 25 would advance to the next round. How many rounds prior to the final round would depend on how many songs are eliminated by the first poll. For example, if the top song on The Black Album (which had no singles) doesn't make the top 25 in the first-round poll, there's no need to poll other songs on that album in a subsequent poll. (I say this as a huge fan of The Black Album--I think it's quite possible that more than one track from that album could make the final top 25, but we'll see.)

I'm just throwing these numbers out without looking at all the polls. But the idea is to choose a group of songs that came close to making the top 25 by some measure and then pick the best from that group to add to the final poll. Sort of like a sports tournament with multiple elimination rounds. If the final poll includes more than 100 songs, we aren't going to get a lot of voters because it will take too long to complete, especially for orgers who aren't intimately familiar with all the albums. Imagine trying to compare 100 songs in one poll from different phases of Prince's long career, some of them unreleased--not an easy task! You want each poll to include a manageable number of songs. On the other hand, we have to keep in mind that some of the polls didn't get a lot of votes and may not be that accurate. The A-sides poll, for example, doesn't always match the ordering of songs in the individual album polls with respect to A-sides from each album.

I don't see any harm in repeating the SOTT poll. We could include the top songs from the original poll along with the top songs from the revote, so that people who voted either time would have their voice heard. When I did that poll, people weren't familiar with the rules so I had to end it early before it turned into a fight between those who like "The Cross" and those who don't.

One thing I would say for sure is that the final poll should only be conducted once, it should be advertised well in advance (maybe made a sticky), and it should be kept open for longer than the other polls if possible. Once that final poll has been conducted it will influence people's thinking one way or the other, so we need to get it right the first time.

I totally agree with this. I had also thought about the sticky and the longer duration. As people have become familiar with the norms, I don't think it'll cause problems to have the poll open for, say, four days.

This methodology is quite interesting also in that it allows to rank a great number of songs while, at the same time, having an affordable final poll. I mean, the songs that are eliminated in the, say, third round (just a guess) will turn out ranked and could be added to the results of the final 25 songs-poll. So could we have a big list and, at the same time, polls with not too many songs, so that the number of voters is high.

If the mods make the final poll thread a sticky, maybe they could lock the thread until the poll closes so that people can't report the results there while the poll is open. After the poll is closed we can all go to town.

I think that, as UltraLash has suggested, we should make use of all the data from the earlier polls in devising the final polling framework, so that everyone who voted in each poll (including the album poll) will have had some impact on the composition of the final poll. That is why I like UltraLash's idea about including more songs from the albums that did well in the album poll. I wonder, however, if the albums with two or three discs (like the one we're voting on now) should get a couple of extra entrants in the first round to account for the larger number of songs on those albums.

If we have enough rounds of polling, the various sampling errors in the individual polls should be ironed out so that everyone (or almost everyone) can agree that the songs in the final poll really deserve to be in there and the other songs were justly eliminated. There will be enough to criticize in the results without people questioning the methodology.

Up to 36 votes! Keep going!

I'm going to bed now. 42 votes. Still many hours left to have your voice heard!

The poll is now closed. Crystal Ball has won very easily. Sad showing for my favourite tune, Last Heart: number 10.

There has been an interesting fight between some outtakes from the Gold Experience era. Finally, Acknowledge Me has prevailed over Days Of Wild and Interactive.

The next poll will be The Rainbow Children. Don't miss it!

Javi said:

The poll is now closed. Crystal Ball has won very easily. Sad showing for my favourite tune, Last Heart: number 10.

There has been an interesting fight between some outtakes from the Gold Experience era. Finally, Acknowledge Me has prevailed over Days Of Wild and Interactive.

The next poll will be The Rainbow Children. Don't miss it!

I definitely think that at least the top 3 in this poll should automatically be entered in the first round of final polling. The problem with this set is that it ranges widely from some of his best songs to some awfully weird stuff.

electricberet said:

Javi said:

The poll is now closed. Crystal Ball has won very easily. Sad showing for my favourite tune, Last Heart: number 10.

There has been an interesting fight between some outtakes from the Gold Experience era. Finally, Acknowledge Me has prevailed over Days Of Wild and Interactive.

The next poll will be The Rainbow Children. Don't miss it!

I definitely think that at least the top 3 in this poll should automatically be entered in the first round of final polling. The problem with this set is that it ranges widely from some of his best songs to some awfully weird stuff.

Hmm. By my fancy pants formula, only the first two songs would be eligible from this particular album -- Crystal Ball and Crucial -- which would be fine by me. I don't see any tracks from this album ending up in Prince's top 25 anyway, although I do like the top two.

UltraLash said:

electricberet said:

I definitely think that at least the top 3 in this poll should automatically be entered in the first round of final polling. The problem with this set is that it ranges widely from some of his best songs to some awfully weird stuff.

Hmm. By my fancy pants formula, only the first two songs would be eligible from this particular album -- Crystal Ball and Crucial -- which would be fine by me. I don't see any tracks from this album ending up in Prince's top 25 anyway, although I do like the top two.

I guess I'm just a sucker in the dream factory.