This takes Ehler's work that dynamically finds the frequency of market cycles and applies it to the concept of Bollinger Bands.
First the dominant cycle length is found using Ehler's methods, this is then used as the length to the moving average and deviation.
I also ended up using an exponential average rather than a simple average for both the moving average center line and the sum of the squares when finding the deviation.
credit goes to LazyBear for coding Ehler's original ideas in pine, which I then used for this script.