Creating set lists. Whew! what a load of work. In a few days I’ll bust out a set of live music, but I wondered (to myself, but still loud enough to confuse the people sitting beside me on the train) how do I decide on song order?
This can’t be simple. That just won’t do. So let’s overthink it.
At first, I simply ranked the songs’ “goodness” but it was not enough to create order among the 14 tracks. So I added a few more categories: tempo, mood, dynamics, melody, lyrics, and playability—from those I derived the average score and the median score for each song. My initial plan was to use the average and/or median scores to determine which order I should play the songs in. But those metrics weren’t as descriptive as I’d hoped. So I decided to create weighted averages for the dynamics, tempo, and playability scores; my logic being that these three metrics somehow represented the three most memorable characteristics of a song.
So I have all these numbers, but what was I going to actually do with them? How was I suppose to arrange a list of numbers to translate into a cohesive and enjoyable set of live music? I began by plotting the numbers with the songs in a pretty random order. The graph did not look like a road that I wanted to travel on. And that pretty much became my standard for a good set list: would I want to drive a car across the graph of its associated scores?
I finally decided to choose “tempo weighted” as the most crucial metric. I did this mainly because the variance in these scores was highest, i.e., there were higher peaks and lower valleys than in other categories.
Next, I needed to decide on how i wanted my “road” to actually look. I chose to start the set off mid-to-high tempo and sustain for a couple of songs. Then slow things down and then back up to mid-tempo, and then slow things down again. The set would end with a climb upward to the highest tempo song. Here is the result:
All in all, I think it’s close to arbitrary but it always feels nice to have some reason behind “big” decisions like this.