I had looked into this ten years back as I always look at all the bad in life first, it seemed like a nightmare to myself and attorney, as so many ways of being bent over even though you trying to be fully in compliance. I believe the SEC is much tougher on regulations and they should be to a point, but also they need to look within as Madoff got away for years. Glad I didn't go that way and concentrated on own funds.
1-It depends on what asset class you are trading, where your place of business is and where your clients are. You will be set up as the authorized trader and be given limited authority over the account. In the USA, for futures, you need to look at CTA exemptions and rules. For securities, look to your state regulator to see if you need to be an RIA. 2-Have clean contracts and meet any regulatory requirements. 3-Most managers are LLCs.
Not true actually. You don't tell C2 how your strategy works. All subscribers get are signals. C2 does have a coding language platform that you can use, if you prefer, to build an automated strategy. It is hosted on C2, but subscribers won't have access to it, of course. Otherwise, you simply enter your signals/trades into the C2 website (or other means); and the signals are sent to subscribers. Some subscribers set their brokerage accounts to auto-trade the signals. Or, you trade your own brokerage account as you normally do, and your trades are mirrored to your subscribers, real-time. (They also have API's, etc.)
I'm quite interested in how such bots work. And, most importantly, what glitches are possible within them? Would be really sad to lose all your money because of a mere bug...
Although automated systems can cause terrible damage in wrong hands the answer would depend on many factors. Some of those are: Choice of programming language. There are languages that provide complete control of the execution, and others that run in a protected environment. I chose letter one to eliminate a possibility of data corruption. Testing. I have unit tests that mock various market and account conditions and validate strategy output. This verification is executed after every single change to the code. Type of strategy. Some bugs may still slip through unit tests, regardless how much it is tested. If the strategy is leveraged a wrong decision may cause significant loss. I decided not to go with such strategies. Instead I'm trading major ETFs and options, and before every single trade is made the system validates account value, existing positions and the order to not to exceed allocated amount. So in case of an error I will notice a performance degradation, but would never loose everything or a significant portion at once.