I gave it a half-ass try and got some results that look like it might have some potential with tweaking. If it turns into anything, I will let you know.
Also, when trying to reproduce the strategy, I had to use an absolute spread (difference) between the two instruments and not the ratio. I also normalized the prices before calculation. Sound about right?
In my tests, I never managed to get great results using Bollinger Bands...
I should try more, though
I gave it a half-ass try and got some results that look like it might have some potential with tweaking. If it turns into anything, I will let you know.
Also, when trying to reproduce the strategy, I had to use an absolute spread (difference) between the two instruments and not the ratio. I also normalized the prices before calculation. Sound about right?