Creating own trading system

hftvol, how about studying a master of financial engineering? Would that help making a profitable trading system?

 

Hong Kong stock exchange is much worse than US, they are so loose and hope foreign giant financial powers stay in HK. Basically it is an open casino and full of extreme not fair products for open to buy. Even those giants did anything wrong, HK side wouldn't punish them at all. Definitely no serious control of giants like what happen in US.

 

GETCO has offices in Chicago, New York, London, and Singapore. On March 21, 2013, the company announced the closure of its Hong Kong office, due to Hong Kong's slow adoption of high frequency trading. On the same day, GETCO received regulatory approval to start operations in Mumbai, India, becoming the first western trading firm to tap into the country’s fast-growing electronic derivatives markets

http://www.marketswiki.com/mwiki/Global_Electronic_Trading_Company_(GETCO)

 

Spot on. BTW, the combination of this knowlege - expert C++, CLR, systems, etc. and intimate understanding of markets and complex products is incredibly hard to attain and thus very rare. Anyone with a lesser amount of knowledge and drive, who happens to think they can buy a cut of tradestation, multicharts or whatever and code up five "winning" strats in EasyLanguage in order to compete with said expert, and trade their way to even a mediocre living, is a complete lunatic.

/Wulfrede

 

Was the story of turtle trader a complete lunatic novel story?

 

Turtle traders were not day traders and certainly were not involved in anything even resembling HFT.

I actually am in favor of longer term trading, where automation plays a smaller role. But that implies that one is very well capitalized and doesn't mind huge drawdowns. I made a point on some other thread that successful daytraders eventually either diversify or transition to longer time frames in order to improve their E at the cost of higher equity vol and capital commitment. There is a merit to turtles and their likes but it has nothing to do with my argument.

/Wulfrede

 

A modified version of turtle system may works in day-trade range? There is future like mini s&p that trade 24 hours a day, there is no different between day-trady or swing trade for future like this as long as not trading crazy fast like HFT...

 

How to put this plainly... it doesn't.

No system "works" on any timeframe. You might as well learn this early on. The whole argument that a given system (as opposed to a method) applies to all timeframes because markets are fractal is complete bullshit. The markets *look* like they can be traded the same way on all timeframes, but nothing could be further from the truth.

What does work is a set of underlying factors prevalent in their particular timeframes that drive the price. All these waves, top and bottom formations are largely random in themselves, but they are created because of non-random overarching reasons that must be well understood before you try to buy something at "support". And no, this is not a contradiction.

I just went and had my system generate a chart of "GOOG" data that looks tradeable. You see tops, bottoms trends, support, etc. I promise you, the whole thing is completely random (well, pseudo-random). Go ahead and turtle-trade it. The reason you can't is that the price action shown on this chart isn't driven by anything in particular. But sure looks like it is, doesn't it?

/Wulfrede

 

So you mean base on your knowledge, turtle system contains some good points that are the drives in many days kind of trade, but the day trade level "reasons" that create the raise/drop is totally different so turtle system wouldn't work right?

Base on what I read, Ed Seykota said that he started with modifying Donchian system, and turtle system is a modified version of Donchian system. Seem like this system in general tells some basic factor of raise/drop trend in a "many days" time frame.

 

Ed Seykota said day trade is not an effective way to win money for an individual mainly because brokerage fee/profit or lose is too hight comparing to overnight kind, and it implied there is like no way to predict for day trade range movement. How do you guys think about this?

HFT definitely shows it is possible to win from day-trade, but companies with HFt normally has super amount of money that can affect the market and has almost zero kind of transaction fee. Seykota definitely was not talking about case when an organization has that resource.