Good introductory course in Algo trading!
Being a computer science person and am doing a master degree in investment management at HKUST (going to be my 4th degree), I've always wondered what the buzz was about Algo/high frequency trading, particularly the details behind. This was a good introductory course to the details I have been wondering about. Ran into this class by accident, initially thought I would just sit in for a couple of lectures to close some of my curosity and end up sitting through the whole course lectures.
Doctor Haksun gave us very good introductory details on some of the popular trading strategies (hidden Markov, mean reversion/pairs trading, trend following/technical analysis) as well as the key mathmatics behind the strategies. I consider myself a pretty good number/mathematics person but I wish I had more background in statistics/mathematics though so I understand more of the mathematical derivations behind the scene. Nonetheless, I got what I was looking for, to better understand what the buzz was about and the key details behind it.
Having gone through almost all the class materials (one more class to go), I remain to be convinced about how viable this is though. That is, with what we learnt so far, can we viably implement something that would make statistically meaningful positive return using that alone? The key block I have not been able to move pass still remain that we are still using historical data (asset movement pattern, returns, risks, etc) to predict the future, which likely is not viable/accurate. But then, I do not have a better alternative to model this, not yet anyway, so the thinking/research continues...
I really wish there was more time though, perhaps this should be a 2 course curriculum: 1 to teach the basic and the theory and then a follow on class for us to implment. Amount of time is very short for the amount of materials and types of materials that we have gone through. I really would have like to fully implement one of the trading strategies, back test it, and then add some of my own "improvements" on the top of the basic and then freeze the code and simulate how that work on future 2 months data for example, afterward evaluate how we do. That would have been the major highlight of the class I think, at least for me anyway. This one month course was simply too short for that...
Doctor Haksun is pretty funny and (smart) obviously very good at what he does. I am glad we get to meet.
Thanks for the wonderful class and spending the last month with me.