May 3rd, 2014, 10:36 am
QuoteOriginally posted by: farmerThese big trend traders don't have a correlation problem, or a math problem. They have a low-frequency kind of problem. There may be a lot of competition, and they may have taken on a lot of capital, but their main problem is that prices haven't made big moves lately.Right.QuoteSo they are pushing into higher-frequency, order management, better chop filtering to find the smaller trends sooner. What would really help, is some way to say something like Europe is not moving into a new regime, China is. Therefore, moves in the yuan are likely to travel, moves in the euro are not.So how do you quantitatively decide, in an automated way, which trends to overweight like this? One idea, is to go beyond time-series data, into text data. Machine orders go up, machine orders go down. But suddenly there is a new word appearing twice as often as "machine orders."No. You cannot predict the future.When there are no trends, you sit tight. You keep doing your thing, and your account goes sideways.What most people entering finance lack is a willingness to sit tight and do nothing.You cannot do that if you are working for someone, including if you are a one person hedge fund trading clients money...But you can do it if you are trading your own account.