Vega, I've tried to stay flat. Its gone from -2000 to 5000. 30/70 split between negative/positive delta on vega distribution. Theta, when I first started out, I was always selling premium and trying to get 1 or 2% of portfolio value, So around 2k to 5k positive theta. Then after reading Nassim Taleb's dynamic hedging, discovered the magic of long options. Theta has been averaging 500 to -1500 ever since. I try not to never let this go below -2000. Again this all depends on market conditions. If I think of a consolidation or bearish bias, these numbers may flip. On the subject of algos giving us high ROIs, the few mainstream hedge-funds who have been doing triple digit returns through algos and HFT are Renaissance and Citadel. I tried to put my money at Renaissance and I was told (through a referral) that I needed a minimum of 10M or be an employee of the company. My question is whether a retail trader such as you and me can algorithmize any potential high profit strategy to reach high double digit growths?
I probably should have mentioned this, but I am currently building my own system. So I can speak from experience. FYI: I have been consistently working on it for over 2 years now (like I said, non-trivial undertaking). I vastly underestimated the amount of work involved at the start and probably would not have continued if it was not for the fact that I am also using it as a learning experience. Its been rewarding in many ways. Its still not in a state where I can use it for trading and maybe it will never be what I hope it will be. But for now its more about the journey then the goal
Hey, same here. I have a strict policy to NOT add new functionality unless it has potential to pay for itself
If you parse the raw feed it probably is. The cost to use a direct feed is also very high, better off starting out with a redistributor like IQFeed. That is what I will probably do.
>> Not sure how it's less expensive to build everything from scratch. Well I'm one of those latency insensitive guys which means I don't give a shit about sub-nanosecond, edge of Paretto edge of computing science crap that high speed automated recruitment bots put for ads in automated no brains if you got the muscle big bank jobs. But I still care about getting an execution today rather than tomorrow, which is the #1 thing I'm doing this thing myself, the other being I'm monitoring more than just one index, and at even 3 split per continents it becomes tedious for the human eye to keep the pace.
>> the few mainstream hedge-funds who have been doing triple digit returns through algos and HFT are Renaissance and Citadel I interviewed at Citadel and was underimpressed (to say the least) by the depth of wits of the hiring manager. Which, given what I know about how all this trickles down, leaves us with perhaps, only one contender in the market. If you're willing to make that bet and go for the best there is
I don't trade options but I think most ligit brokers have good execution nowadays. Getting market data to your strategies in a timely manner, that's a challenge. You planning to co-lo?
It almost impossible to find 100% required platform. Each trading style follows individual platform. Try to use some blocks to combine own solution. Lean as said before. StockSharp also. Some non trading libs like protobuf.