--------------------------------------------------------------------------------
Quote from fkeane:
Neiderhoffer and many others, however, claim that financial history is non-stationary. To borrow on one of his analogies, its as if the urn containing thousands of red and white marbles has an elf at the back, replacing the mix of marbles. So when you sample the data to figure out its patterns (what percentage each of white or red marbles in the population) you will be right for a while, but in time the population changes. Furthermore, you will lose money for a while because you wont know if something has really changed or if you are just experiencing normal volatility.
--------------------------------------------------------------------------------
Think about that a little bit. Dennis Richmond blew up twice following periods of great success. The turtles all saw their returns deminish after 4 years. It would seem to me that system traders are especially vunerable to this concept of non-stationary markets, emphasised by the fact that trading these systems neccessites a blind faith during draw down periods that the past market behaviour supporting the system's viability is still persisting.
My fear (as an aspiring system developer) is that this non-stationary attribute of the market is becoming more volitile, giving rise to smaller window periods of profitability for any given system. The huge influx of traders utilising system development (see bottom of this link:
http://www.nypost.com/business/23134.htm) poses a logical rationale supporting that profit opportunites will be identified and exploited away much more quickly than in the decades past where a handful of revoutionary traders like Dennis
were doing the hard slog of back testing manually. The Turtles had 4 years of phenomanal returns, that thereafter disapated somewhat, and this was perhaps not by coincidence concurrent with increasing acceptance of systematic trading methods and more sophisticated and powerful analytical tool (the trusty 'puter).
I would love to hear from Curtis or any other previoulsly successful system traders regarding the impact proliferation of system development has had on potential profitability. Seems to me that the type of trader like Paul Tudor Jones might just have the advantage over the system trader now - for him the market can be as non-stationary as the market likes because of the forward looking approack of reading the macro-environment.
Sam