According to most authorities, stock markets are non-linear,
dynamic systems. Chaos theory is the mathematics of studying just that-
non-linear, dynamic systems. Chaos analysis has determined that
stock market prices are highly random, but with a trend. The amount of
the trend varies from market to market and from time frame to time
frame. A concept involved in chaotic systems is fractals. Fractals
are objects which are "self-similar" in the sense that the individual
parts are related to the whole. A popular example of this is a tree.
While the branches get smaller and smaller, each is similar in structure
to the larger branches and the tree as a whole. Similarly, in market
price action, as you look at monthly, weekly, daily, and intra day bar
charts, the structure has a similar appearance. Just as with natural
objects, as you move in closer and closer, you see more and more
detail. Another characteristic of chaotic markets is called "sensitive
dependence on initial conditions." This is what makes dynamic
market systems so difficult to predict. Because we cannot accurately
describe the current situation band because of errors in the descriptions,
they are hard to find due to the system's overall complexity. Accurate
predictions become impossible. Even if we could predict tomorrow's
stock market change exactly (which we can't), we would still have
zero accuracy trying to predict only twenty days ahead.
A number of thoughtful traders and experts have suggested that those
trading with intra day data such as five-minute bar charts are trading
random noise and thus wasting their time. Over time, they are doomed
to failure by the costs of trading. At the same time these experts say
that longer-term price action is not random. Traders can succeed in
trading from daily or weekly charts if they follow trends. The question
naturally arises how can short-term data be random and longer-term
data not be in the same market? If short-term (random) data accumulates
to form long-term data, wouldn't that also have to be random? As it
turns out, such a paradox can exist. A system can be random in the
short-term and deterministic in the long term.