Tuesday, April 17, 2007

Nerdcore Top 10 update

My head hurts. The next time I get the bright idea to do a collaborative top ten project like this, someone please remind me that I am a man of letters. Numbers frighten me.

This thing could have easily been a clusterfuck. Seriously. No one could’ve participated or everyone could’ve just bribed their friends to flood me with bogus lists designed to make them stand out, but it didn’t go down like that. I actually got a ton of submissions - well over a hundred that I’ve recorded and a growing pile of indeterminate size that I haven’t - fraught with genuine thought and sentiment. A lot have come from fans, thanks to the support of cats like my pal Anthony and the rest of the crew over at Game Music 4 All, and a lot have come from the artists themselves. I’ve been presented with an interesting cross-section of the nerdcore populace, to say the least.

The final list is slowly beginning to emerge from that undulating mass of names and numbers that has so bedeviled me for the better part of a week. But, saying that, I fully realize that the next time I look in the ol’ Inbox there could be a handful of submissions that completely change the terrain. I think that’s the part I like best.

The part I like least is the whole process of quantification. While a lot of people assigned a specific position value to their preferred artists, many chose not to. I’m totally cool with that, of course, as I understand that one musician’s merit isn’t always so easily contrasted against another. Still, this presents an interesting quandary; do I focus on an artist’s overall position or on his sheer prevalence?

My answer: both.

Basically, in addition to keeping a running tab of how many times a specific artist is mentioned, I also keep a record of this artist’s average position within weighted submissions. The plan is to reconcile the results via a concordance of both lists, as this seems both the fairest and the most accurate way to gauge the results.

So far it looks like this should be eminently doable, as the results so far seem to jive between both lists (meaning that those most voted for are also generally the cats most highly ranked).

Does this mean the list will please everyone?

Oh fuck no!

But, thankfully, that was never the intention of the project. :D


ChurchHatesTucker said...

Interesting, Z. I'm trying to figure out how I would do it, and nothing seems quite satisfactory.

Z. said...

Welcome to my world, Church. ;)

ChurchHatesTucker said...

OK, here's what I would do. Seems obvious in retrospect.

Assign points for each position in a weighted list. Ten points for first place, down to one point for first (a total of 55 points for each list.) If someone has two people tied for a position, then they would split the total of those two positions (the one they're tied for and the one that's left vacant as a result.) Thus, a tie for first means that each gets half of both first (ten points) and second (nine points) places, or 9.5 points.

Then, all you need do is treat unweighted lists as ten-way ties. Each entry is worth 5.5 points (one tenth of the total points of the list.)

Add them all up, see who's the top ten and you're done.

(I encourage any math geeks to point out flaws I've overlooked.)

ChurchHatesTucker said...

Argh, should be "Ten points for first place, down to one point for tenth."

antisocial said...

Sounds reasonable enough. The only possible flaw (and I really can't conceptualize it enough to tell if it is a flaw or not), is that when people have an unweighted list, certain artists may get bumped up higher than expected in the rankings. For example, say there are a bunch people and I get to pick three of them. I pick:

1) Rob
2) John
3) Bill

A second person picks:


A third person picks (in no particular order):

2) Sam
2) Bill
2) Steve

Rob - 3
John - 2
Bill - 4
Mike - 3
Steve - 3
Sam - 2
Gary - 2

Although nobody voted for Bill as 1, he still gets bumped to the number 1 spot because of the person who didn't order their list (e.g., me), and thus elevated his position on the list. This is probably not a good example, because Bill got 3 votes, while the other people only got one vote.

Meh, I quit. My head hurts.

Z. said...

I thought of doing that, Church, but then I struck upon the very problem Soc speaks of. The bottom line here is that I wanted to establish a consensus. That’s why I didn’t simply have a nomination period and then just let people vote for folks on some master list. Essentially, I’m using the number of times an artist is mentioned to narrow the filed down to those most admired, and then I’m using individual artist data concerning average positioning to determine his place in the final list. I like this because, at present, the number of votes folks have are almost always inversely proportional to their average relative position.

ChurchHatesTucker said...

Soc, remember that you're dealing with unweighted lists there, so 'Bill' could just as easily have been listed first (or last) on each.

And even with weighted lists, that result is valid. Remember that the Tour de France winners seldom win individual stages of the race.

Beefy said...

um...fuck math. that is all.

Tetsuo said...

"fuck math"? "fuck math"? And you call yourself a nerd! Sheesh, I don't know, some people...

(note: post may contain elements of non-seriousness)

Steve said...

I'm glad you're not calculating my SAT score, Z.

Steffo said...

Physical Challenge!!!

Z. said...

I would pay good money to see MPFM his the Obstacle Course. :D

ChurchHatesTucker said...

I can totally plot out Steffo's obstacle course performance!

OK, what are the percentage of aerobic activities versus strength-based activites? Is dexterity involved? Does anyone have a d72 handy?

Antisocial said...

I remember one time I was playing a game in which 2 people had to pick a number 1 through 6.

I spent about three hours arguing that they had a 1 in 6 chance (as opposed to 1 in 36 chance) of being the same number.

Against statistics majors.

That was a sad day.

Thanks for setting me straight, church.