Hi Link -- Thanks for the kind words about my book. There is another -- Introduction to AmiBroker -- that has a set of tutorial exercises that might also be helpful.
You will have read in QTS my thoughts about the markets being very nearly efficient. If a large inefficiency existed, and we shared that information and designed trading systems to profit from it, each trade made would reduce the inefficiency by some amount. Eventually the inefficiency would equal the frictional costs of trading the system -- the commissions, slippage, taxes, office expenses, and so forth -- and trading it would be profitable only to those traders who were able to best anticipate the signals, had the best execution, and had the lowest costs. My view is that it is difficult to design a profitable trading system. Most of the ideas that are easy to try -- such as breakouts based on end-of-day data trading at the market on the next open -- are tried by many people on a regular basis. If one of those systems tests as profitable and begins to signal profitable trades, the people investigating it will trade and remove that inefficiency very quickly. So, your job as a system designer is to identify an inefficiency that either: 1. most of the trading world does not know about. Then you can trade it profitably until the inefficiency it recognized disappears -- either naturally as markets shift characteristics, or as more traders use systems that recognize that same inefficiency and collectively you remove it. 2. exists for some fundamental reason for some period of time, until the fundamentals change. My advice is to become very fluent in the use of AmiBroker and AFL, so that you will be able to easily program any system you can imagine might work. (There is no way to avoid becoming a fluent AFL programmer. Just accept that and do the work necessary to gain that skill.) By all means, practice using systems that you read about or find in a library. The AmiBroker Library is an excellent place to start. Work through as many of the listings as you can. Learn from both the good and the bad. Play with all the features of AmiBroker. Build your experience. Then think beyond the systems described in easily obtainable publications -- whether those are books, magazines, programs, or blogs. Be willing to use systems that are more complex -- that use more advanced math or cleverer logic. Be willing to use shorter time period bars to make your decisions. Pay particular attention to the issues you trade and the risk you exposure yourself to. Be sure to use good system validation techniques, including in-sample and out-of-sample testing and walk forward validation. You already know the importance of selecting the objective function that fits your trading situation, personality, and psychology. Do Not begin serious walk forward testing until you have selected Your objective function. There are many people posting to this list who are trading profitably based on mechanical systems. And probably many more lurking in hopes that someone will give them some good ideas. Thanks for listening, Howard www.blueowlpress.com On Tue, Jan 13, 2009 at 12:43 PM, whfh5 <[email protected]> wrote: > Hello everyone. I'm brand new to AmiBroker and trading system > development. I am reading Dr. Bandy's book, Quantitative Trading > Systems and have found it very helpful. I was going to get setup with > TradeStation until I discovered AmiBroker. My first qusestion is, how > do I go about developing a trading system that works? Do I just pick > random indicators or test those I read about? According to Dr. Bandy, > any decent system that has been in play for a while eventually becomes > non-profitable. I can imagine no one with a working system is going to > want to share it? How many people here are actually trading with a > system that beats the market? How long to systems typically last > before the fail? Who do you recommend for data feed? > > I'm sure I'll have more questions and thank you in advance for the > help to a newbie! > > Link > > >
