Numbers!

Hi Everybody,

So, last night I got the urge to get out the database and look at what's been going on so far. This was partly inspired by a desire to go a bit deeper with some of the data used for the "Case For Support" post.

I am, of course, aware that statistics are often used to lie and distort information - but I do think that they can be a useful tool if you know what you're looking at. I will also say that I'm not a statistician, nor am I intimately familiar with the science of statistics in general. However, I do know just enough that some interesting meanings might be extracted from numerical descriptions of what happens at Underground.

Also, it is very important for me to state that any "negative" connotation of terms used in this post is only towards the numbers themselves. It's not a reflection on anyone's show, or their promotional efforts. Indeed, having spoken to a number of folks, I think I have fairly strong anecdotal evidence that "general public" promotion has no appreciable effect on show turnouts - so, save your money, and just tell all your friends to bring their friends. Flyers don't work!

First off, the total number of shows so far this year:

111

...and the total number of attendees (this is derived from the door split, and so is not 100% accurate - but it is close enough for rock, and/ or possibly roll):

4112

If we divide the total attendees by the number of shows, we get an average nightly turnout of:

37

For the purposes of looking at things further, I think it's wise to establish a "window" for average numbers. That is, it's probably pretty rare that EXACTLY 37 people show up at an event, so we need a bit of wiggle room to get a meaningful count of how many shows were "extremely average." For the purposes of this post, I'm going to define a 5% "swing" around the exact average as still being within average "territory." So, for these numbers, an "average show" might have a range of attendees between:

35 (low side) and 39 (high side)

Moving on, if all the raw turnout data is handed off to a "Standar Deviation, Entire Population" calculator, we get a standard deviation of:

30 people

In my mind, what this means is that the turnout for a "reasonably normal" show has quite a range of possibility. You could have anywhere between 7 and 67 people in attendance, and still have a show that would be reasonably considered to be "ordinary" in terms of turnout.

Now then, if we do some counting, we find the following:

"Average" Shows: 9

"Above Average" Shows: 43

"Below Average" Shows: 59

What's interesting about this is, that if you look at how the turnouts "map" to the different overall show divisions, you see that 39% of the shows (a minority) accounted for 68% of the total turnout this year so far. In fact, if you add up the attendees for shows that were average or below, the total number of people is 1324, which is still less than the number of attendees of shows that had turnouts above average (2788 people).

Let's go a little deeper:

"Uncommonly Small" Turnouts (below 7 people): 10

"Uncommonly Large" Turnouts (between 67 and 97 people): 9

"Rarely This Large" Turnouts (between 97 and 127 people): 4

"Exceptionally Large" Turnouts (more than 127 people): 2

"Normal" Turnouts: 86

This is interesting, because it means that only 14% of all the shows done this year had 35% of the turnout (or, in other terms, 15 "better than normal" shows accounted for 1458 people that showed up).

Fascinating, eh?