Commentary, Pt. 2: Pitchfork Top 500 of the 2000s

Lists like the Top 500 Songs of the 2000s are great because they offer a jumping-off point for a conversation that could last days. Since these lists are inherently subjective, and since I have a simple engineer's brain, finding objective trends in that sea of personal opinion is always fun. We saw yesterday that, based on sheer number of songs in the Top 500, the order of years in terms of "quality music" is as follows:

1. 2003 (63)
   
2. 2005 (61)
   
3. 2002 (60)
   
4. 2000 (59)
   
5. 2006 (58)
   
6. 2001 (52)
   
7. 2004 (51)
   
8. 2007 (48)
   
9. 2008 (33)
   
10. 2009 (15)
    

Clearly, that's a blunt instrument for measuring quality, so the question becomes how to best analyze the rankings. Maybe average rank of songs from each year is the best mechanism? Maybe the median? Not surprisingly, things change significantly when you look at statistics like that. Using the average rank (the median was essentially the same) for songs released in a given year, the hierarchy of "quality music" then becomes:

1. 2007
2. 2001
3. 2005
4. 2000
5. 2003
6. 2002
7. 2006
8. 2004
9. 2008
10. 2009

What would cause such a large-scale change in the order of years? A cynic (or statistician) would say that the smaller sample size allows for a higher ranking, and there is undoubtedly some sample bias inherent in the calculation. Or maybe music from 2007 was just better? But we already said that ranking 500 songs is ludicrous. What would actually qualify as "the best music?" The top 250? 100? How about 50? If all years were equal, each would likely have have 5 songs in the Top 50? Maybe a good idea would be to analyze the overall rank of the top 5 songs from each year... for example, if the 5 best songs from 2004 were ranked 5, 10, 15, 20, and 25 overall, the Average Top 5 ranking for 2004 would be 15. Well, when we ran the numbers for the past decade, our Average Top 5 ranking comes out like this:

1. 2003 (13)
2. 2000 (14)
3. 2007 (17)
4. 2002 (18)
5. 2004 (25)
6. 2005 (27)
7. 2001 (34)
8. 2006 (40)
9. 2008 (69)
10. 2009 (144)

What falls out are some pretty obvious tiers, which in all honesty is probably the best way to look at something like this. Occupying the Top Tier would be 2000, 2002, 2003, and 2007 (interesting, since 3 out of 4 of those years had the most overall songs in the Top 500... we get quantity and quality). 2001, 2004, 2005, and 2006 occupy a Middle Tier; however, I'm not willing to call 2008 and 2009 the "Bottom Tier." Let's just say "Not Yet Appreciated" and leave it at that.

Tomorrow, we'll actually dive into some of the songs from each year, and maybe try to defend and/or argue the inclusion of a few.