Quant development: Experiment of Statistical Advantage Portfolio (S.A.P.) Need your comments.
Hi all, here are some thoughts regarding my Statistical Advantage Portfolio (S.A.P) plan. Please let me know whether i am on the right track. Or just waste my time.
S1. Select trading strategies from Internet with positive return using 2-10 years data for testing.
S2. Convert those strategy to MQL4 or MQL5 program and test them using 2-10 years data on my own platform.
S3. Based on the testing result, pick the most profitable currency pairs and programs.
S4. Put all those programs in one portfolio with 1-2% risk per order.
S5. Review program performance bi-weekly. Increase or reduce risk level based on program performance.
Again, am i on the right track to build an algo trading hedge fund ?
A portfolio of strategies sounds reasonable, as a start. Additionally, I would acknowledge that strategies rotate — some work for a while, then cease to work, then work again. I would investigate the trendiness of strategies and overweight those improving and underweight (or eliminate) those declining. Said a different way, I would have a stable of strategies running on paper, and only those working well would be used in the actual portfolio.
hanks for the feedback. I plan to add in some additional logic to track the performance of every system and adjust there risk level automatically. for example, if the system continuously hits the stop lose for two times on one currency, the risk might be reduced for 25%. on the reverse side, risk level could be improved 25%. that’s the thought.
It is close to the you guys doing in the industry ? or it is too low class 🙂
I can’t say what the industry is doing; I’m just an interested party, not a practitioner.
From a Linked in group discussionNOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!