I like the concept of quant trading, removing subjective bias from trading and putting actions purely on numbers. I've been trying to create a multi-layer system to pull from many sources. The Markov chain was interesting, but I have not found a way to use it as a standalone strategy. My back testing showed it had decent macro state detection but weak micro timing precision. Combined with my strategies it reduced the draw down considerably. This lead to suggestions of adding it as a position sizing modifier. for example: - 100% exposure in strong bull persistence - 50% in uncertain - 25% in bear persistence - 0% only in extreme crash probability Which is a good feature, but returns me to finding the ultimate strategy to time entries. I'd be interested to hear how others are using it? My current set up is developing strategies which run through my testing engine to verify effectiveness. Those that pass the threshold go to a paper trading server. The difficulty as mentioned in a video is overfitting, refining strategies to perform perfectly on back data becomes too specific to those exact conditions.