Hacker News Reader

Comment by tutfbhuf | Hacker News Reader

This question can essentially be simplified to: "Is the stock market truly random?", as randomness is defined as the absence of patterns and TA is all about patterns.

Currently, the random walk hypothesis remains just that - a hypothesis. There is no definitive evidence to confirm that the stock market operates in a random manner.

On the other hand, many believe that there are some long-term patterns and invest huge chunks of money based on that belief. For example, one such belief is that the S&P 500 will continue to have an annual average return of something between 8% - 10% in the decades to come.

If the movements in stock charts, including ETFs like SPY, were truly random, such a pattern could not exist. Consequently, our expected average annual return from the S&P 500 should be 0%. This is because in a genuinely random environment, the likelihood of a rise and a fall would be precisely equal.

View on HN

Well...random with a positive trendline can still be random. The fact that there's a positive trend globally doesn't change the fact that the local behavior is indistinguishable from random noise. The way I tend to think about it is that the deltas day to day are completely random, but shifted up from neutral by just a touch, so that the overall trend is upward but the micro behavior is unpredictable.

If I understand you correctly, you believe that "micro behavior is unpredictable", thereby suggesting that technical analysis (TA) is not possible on shorter timeframes. However, you seem to allow for some applicability of TA on larger timeframes, such as the concept of a "positive trendline", a term that occurs quiet frequently in TA. That's an interesting perspective.

I mean insofar as "market goes up over time", that's been pretty much a guarantee, at least on a large enough timeframe, since what, the 1930s? I wouldn't call that TA though, that's basic observation of a trend over time. TA purports to read the market's "mood" and predict its behavior via a bunch of mathematical indicators, which are about as useful and scientific as pickup artistry.

I think this is where the question of “what is TA really” comes in. I think most people would grant there are underlying patterns but there are just so many factors and they are dynamic. Whether the patterns are discernible or not is another matter but in theory if they exist they are be chartable in some form.

The existence of players that generate high-sharpe consistent returns is a sort of empirical proof that the market is not actually random (although these tend to be relatively short term trades, I.e. < 1hour holding periods, turnover portfolio multiple times a day, etc).

In quantitative finance a stock price is modeled by two elements, a “drift” (the long term expected returns) plus a random walk (the randomness). The drift not being zero does not mean that the process is not random.

Reply on news.ycombinator.com