First post here I think and although I'm not running this algo on quantopian I hoped I'd be able to ask this here. I've been working on a statistical arbitrage, mean reversion pairs trading strategy for some time now and one thing that I don't like is the rigidity of my Z-score entry and exit thresholds. Running several backtests to optimize is all very well but I was wondering if anyone had come across any literature or techniques for dynamically choosing the values. I'm not a mathematician, I've learned to program specifically for this strategy so that's pretty limited too. I thought maybe some sort of transform based on the spread volatility which would allow my entry threshold to increase when volatility is relatively higher and decrease when the volatility is low.
Any suggestions would be appreciated,