I do think these polls should be designed to establish a board 'canon', but I feel different people are at different phases of their musical exploration. Somebody new to the site who's just straying away from mainstream for the first time benefits a lot more from the polls as we've been doing them. But somebody who's already been exploring for years and has already heard all the 'canon' albums benefits more from a game that delves a little deeper. That's why I was hoping to find a way to address both groups at the same time: Keep the 'canon' ones in but also leave room for new discoveries.
I agree going by average points per vote isn't the ideal solution, as it doesn't separate the albums a person doesn't like from the albums a person hasn't heard. Also, it makes it so you're penalizing your 20th favorite album but not your 21st.
I wonder if there's some way to come up with some sort of prediction metric for the percentage of voters who have heard the album, based on things like the AM rank, and sales figures that we can use as a constant divider for scores outside the top ten.
Let me add in my two cents here: Every other decade had 10 albums and songs each year. Why not run this poll the same way? If we want to expand or change the system, we should do it next year for the second 90's poll, since we've already done the 90's under this system.
Just to make it clear: I am fine with 100 albums/songs and am happy to do it that way. I just think it can be improved. If we nominated 150 albums and songs and voted for the top 100 from there, I think we would have a more legitimate top 100. If we want to wait until the next go-around of the decades to implement a change then I understand that. My only question is why wait? I don't see a downside to nominating 150 albums/songs and picking a definitive top 100 from there.
I wonder if there's some way to come up with some sort of prediction metric for the percentage of voters who have heard the album, based on things like the AM rank, and sales figures that we can use as a constant divider for scores outside the top ten.
Album sales would be a poor metric to determine how many AM users have heard an album. An album like Since I Left You didn't even dent the charts, but the majority of people here have heard it.
I think it's safer to assume that the AM rank directly correlates with the amount of people who have heard an album. The only exception would be albums by highly acclaimed/popular artists that bombed with critics, but I still think AM rank is the best method short of simply asking people to list every album they've heard from a given year.
