Answerer: Joseph Wang, Ex-VP Quant - 香港投资银行
Generally, the algorithms that HFT firms are pretty simple. However, there are two catches
- the important thing is usually not the mathematical complexity but speed. “buy low/sell high” is the algorithm. Simple, except that you want to receive the data, do the calculation and send the order in 10 milliseconds.
- a lot of the relatively complex mathematics has to do with backtesting. OK. “buy low/sell high” simple. Now run that algorithm over a stock over the last five years so you see how much money you make. Not so simple, but still not too hard. OK, now run that algorithm over all of the stocks in the NYSE so that you can see which stocks that algo works on. OK, now come up with a statistical argument to see that you aren’t really just seeing patterns that aren’t there.
So it’s simple, but complex at the same time.
The other thing is that “technical analysis” is to “algo trading” what “astrology” is to “astronomy.” The thing that makes algo traders professionals is “backtesting.” A professional algo trader will test their algorithm against old data to see how much their algo would have made, and they use hard numbers to see whether or not they are seeing patterns that aren’t there. Also every professional algo trader I’ve met understand the market microstructure to insane detail so when there is a strategy that works, they can tell you why it works.
People that call themselves “technical analysts” are just like people that read horoscopes in the newspapers. They don’t check to see if they made money in the past, and they also don’t think about why “it works.”
Zero 适合科普金融词汇~
发明者量化-小小梦 老大~我英语太LOW了,我要好好学英语。。。