Some things

Better living through biologically-inspired optimization

News from my exciting life

Well, I've been learning more about investment and genetic programming.

It turns out that there are multiple unanswered questions in the intersection of machine learning and finance/investment:

  • Does technical analysis (the use of historical information to predict and analyze trends) have any real merit? Practitioners and academics (as well as academics and academics) differ on this issue.
  • If there exist trading rules which are likely to give returns greater than that of a simple "buy and hold" strategy, can they be found? If so, how?

The literature on investment and genetic programming shows mixed results and returns.

last updated 2 years ago # history

Comments

Well, buy and hold is over-simplified. You'd get killed if you don't pay attention - for example, a perfectly good company like Tyco went completely to hell when the crooks took over, similarly for MCI. I guess what's unresolved is whether people can out-perform arbitrarily picking stocks.
IMO, it's pretty clear that market ativities are influenced by what's already happened, particularly in the recent past, so technical analysis should work. It's just that nobody's really figured out how. It's funny, though, I have a friend who's very into TA, and he completely called the March 2000 crash a few weeks prior. "Get out now", he told me, "the MACD on the Dow and NASDAQ is set up for a correction". After talking to him, I was actually short QQQ the day the crash started, but I closed that position after about 10 days - pigs get slaughtered and all that.

2 years ago # reply

IKlSBh Wow, it can be truth

8 months ago # reply

Add a comment

you're not logged in