Appreciate the post/points! I'm not an expert in this at all myself. Frankly, all of this is picked up from less than 3 years of various small tests in various investment strategies. I welcome all advice/thoughts on this. As said, this is an experiment truly.
Could you elaborate on the beta profiles part? I can't say I'm very knowledgeable there. Assume I know very little about formal risk measurements.
As far as the rules I referenced, I have a max single loss but have yet to set a stop loss/retirement point. This strategy evolved pretty loosely based on the idea of capitalizing on the volatility and part of the reason it's an experiment is that I'm okay with losing the $500 if it comes down to it. It's still incredibly risky, which I am aware of. Part of the reason I'm hesitant to put in a retirement point is that I can see losing a significant chunk in certain scenarios (this algorithm is not yet automated, though it very well may be soon, which would mitigate this) that would still be less than the long term gains. For example, given the returns, the strategy could still perform well taking occasional hits of say 20% in one day infrequently (read a few times a year). If/when this algorithm is automated, I will certainly be building in stop loss constraints.
Fully agreed on the hypothesis and not counting on them at all, but I think they aren't incredibly unrealistic. I carefully chose two years in the post above because I don't see the volatility lasting much longer than that. Right now the strategy is incredibly liquid, and I don't see that part of it changing given the micro focus. No hold so far has lasted longer than 24h, and when this is more formalized, I see a hold time limit (as a function of loss/gain) being used to keep the lost opportunity cost down. I think the tuning will likely occur mainly over the next month or two, and after that I'll likely either stop or let it run.
Overall, I would still categorize this much more as a personal test than a scientific one. We'll see how formal it gets.
Could you elaborate on the beta profiles part? I can't say I'm very knowledgeable there. Assume I know very little about formal risk measurements.
As far as the rules I referenced, I have a max single loss but have yet to set a stop loss/retirement point. This strategy evolved pretty loosely based on the idea of capitalizing on the volatility and part of the reason it's an experiment is that I'm okay with losing the $500 if it comes down to it. It's still incredibly risky, which I am aware of. Part of the reason I'm hesitant to put in a retirement point is that I can see losing a significant chunk in certain scenarios (this algorithm is not yet automated, though it very well may be soon, which would mitigate this) that would still be less than the long term gains. For example, given the returns, the strategy could still perform well taking occasional hits of say 20% in one day infrequently (read a few times a year). If/when this algorithm is automated, I will certainly be building in stop loss constraints.
Fully agreed on the hypothesis and not counting on them at all, but I think they aren't incredibly unrealistic. I carefully chose two years in the post above because I don't see the volatility lasting much longer than that. Right now the strategy is incredibly liquid, and I don't see that part of it changing given the micro focus. No hold so far has lasted longer than 24h, and when this is more formalized, I see a hold time limit (as a function of loss/gain) being used to keep the lost opportunity cost down. I think the tuning will likely occur mainly over the next month or two, and after that I'll likely either stop or let it run.
Overall, I would still categorize this much more as a personal test than a scientific one. We'll see how formal it gets.