I manually execute my trading strategies. Because: I think I can outperform my strategies on average. However this out-performance is accompanied with bigger DD compared to if I execute my strategies using a computer. It also leads to stress and less time for further research. Acrary (ETs undisputed leader in strategy development) has a similar philosophy in which he used to trade 'against' his strategies. As I understand, he would front-run/slightly modify his entries and exits. However, now I am seriously thinking of just using computers to execute for me. Such a step should give me lot more time on my hand for further development and in the end possibly more strategies, and in the long run a healthier & better quality of life. I am curious about the experiences of other systemic and career traders? Have you automated your systematic strategies? Did automation help or not? Have you wanted to automate, but could not automate due to technical limitations? Please contribute your real experiences only. At this stage in my trading career, this is a very important question/choice for me and I appreciate if people can share real experiences. Thanks.
I'm a little confused by your terminology, when you say 'manually execute', do you refer to the process of getting orders filled or the entire process (signals & execution)?
I am referring to entire process - signals, taking 80% of signals as a computer would take them and using discretion on 20% of signal entries and signal exits, depending on context etc. (basically the things not covered in the code yet). I am also referring to changing position size by discretion. So, might have back-tested using 'n' number of contracts, but depending on context, execute n or n+1 or n-1 contracts. Many a times, results are better than backtests, but drawdowns are larger and in fast moving market, price received due to punching orders is 1-2 ticks worse than what would be achievable if a computer was putting orders automatically.
Ok, I see. In response to your question, I automated everything, in fact I don't think I could do what I do manually. I'm monitoring multiple stat-arb positions from many strategies running in parallel. I'm sure if I tried to do this all manually I'd end up making a bunch of mistakes. Sometimes, when reviewing trades, I see sub-optimal entries resulting from the automation, in response to this I will try and tweak the code to improve matters and re-test. I see it as a trade off between slightly sub-optimal automated entries and being able to run a lot of stuff in parallel. I chose the running lots of strategies path due to the diversification it offers. I think, in general, most good automated strategies are hard to trade manually, and vice versa. For me, automating the *right* strategies was the key to profitability. But getting the software to the point of being useful was a non-trivial exercise which took years (and is still an on going process), for me none of the off the shelf software solutions did what I wanted.