The ideal thing would be to have people go to every contra dance for a year and log what figures people danced. How many dances have a balance? A long lines? While this is impractical, it illustrates what we're actually interested in: the frequency of moves at dances. One way to approximate this would be to look at a caller's repertoire. While you don't call each of your dances the same amount, and one caller isn't completely representative of the whole, it should be about right.

Several years ago Rich Goss sent his dance cards out to a mailing list I was on. At the time this was 446 dances. Being both comprehensive and digital, it seemed a good place to start.

I converted the dance cards from a Word document to text, cleaned them up with a mixture of automated and manual tweaks, and then ran them through a program to count figure frequencies. Counting the fraction of dances that contain each figure, and excluding figures that showed up in less than 2% of dances, I got:

It's not surprising that nearly all dances contain a swing. If you call a dance with no partner swing, let alone no swings at all, dancers are going to be very unhappy. I was surprised heys were so common and petronellas were so uncommon.