Music DB Metric

The Music Database Introduction has a table with records broken down into rough stylistic categories. Most of the table lines have something that looks like this (note that numbers in this file are not dynamically updated):

Bluegrass (46) DB (351) [37%]

The interpretation of this is that the database has 351 bluegrass records listed, but I have only rated 46 of these. The 37% number isn't obviously related to these figures, but it is an attempt to measure how extensively I've listened to the category. It is a weighted measure of artists reviewied, derived as follows:

I think the actual numbers tend to skew toward 50%. What I think happens is that the database tends to be thin for categories that I have relatively little taste for and experience in, while it tends to be extensive and even obscurantist for categories that I like and am relatively export in. The former effect is very pronounced in world music, where I'm quite certain that I have come nowhere close to 40% of the major Latin American artists, nor (even more ridiculously) 35% of South and East Asia.

On the other hand, the high numbers tend to max out around 75%. To see what this means (i.e., what I'm missing), let's take a look at Mostly 1940's/1950's Jazz, where we have 191 rated artists vs. 126 unrated artists, including 21 fully weighted: Charlie Byrd, Kenny Drew, Maynard Ferguson, Vince Guaraldi, Ramsey Lewis, Humphrey Lyttelton, Herbie Mann, Big Jay McNeely, Turk Murphy, Papa Bue's Viking Jazz Band, Cecil Payne, Oscar Pettiford, Andre Previn, Buddy Rich, Howard Rumsey, George Shearing, Johnny Hammond Smith, Ira Sullivan, Billy Taylor, Cal Tjader, Charlie Ventura.

For Blues, we have 180 rated artists vs. 195 unrated artists, but the latter are overwhelmingly weighted down, leaving only 11 fully weighted unrated artists: Luther Allison, Ronnie Earl, John Hammond Jr, Clifford Hayes, Earl King, Alexis Korner, Lazy Lester, Louisiana Red, Memphis Jug Band, Roomful of Blues, Clara Smith. In both of these cases, I'm tempted to argue that my coverage is pretty thorough.

On the other hand, Mostly 1960's Rock has 79 fully weighted unrated artists, including: Blood Sweat & Tears, Booker T & the MGs, Tim Buckley, Cher, Chicago, Cream, Deep Purple, John Denver, Neal Diamond, Genesis, Isaac Hayes, Dan Hicks, Hollies, Jefferson Airplane, King Crimson, Gladys Knight, Love, Monkees, Moody Blues, Ted Nugent, Procul Harum, Lou Rawls, Righteous Brothers, Searchers, Simon and Garfunkel, Sir Douglas Quintet, Small Faces, Steppenwolf, Turtles, Junior Walker, Zombies. These are almost all artists that I know well enough to have definite opinions of, but they're unrated here for lack of listening to a relatively recent (or memorable) CD. (Also some notable omissions among the not fully weighted, including: Lovin' Spoonful, Del Shannon, Sonny & Cher, Them.) So I think this low (46%) metric is an aberration.

One more example: Music DB Introduction