I came upon this experiment while learning some Octave modelling language. There was a research done before to see how well an analyst can distinguish a true market price chart from a randomly generated one. The question was whether an experienced chartist can truly see market patterns or his insight was good as a monkey throwing a dart.
I wanted to experiment this with some friends of mine that are big into stocks and technical analysis and thought what they like to say about a randomly generated chart. I made sure to also ask how confident they were in the prediction.
To generate this graph I decided to use the random walk model since I just learned it and was excited to see how it might look in practice. A random walk model, in its simplest form is laid out as such:
where is the price of the underlying asset at time , being the drift constant which we have set to 0 and is an error noise term such that . The reason it follows a studentized t distribution is due to the fat-tails that many financial securities have known to show. The Error term has a standard deviation of 0.05 and mean of 0. Since stocks on the market cannot be $0 or negative, I decided to give it an initial value of
To sample the random t, I decided to put octave into use with the trnd() function. I generated about 500 return periods and put them into a spreadsheet and sorted into 4-period candle charts. The chart is below:
The company name is Random Walking holdings but of course, you don't need to know that :). I sent off this chart along with the tale of an elusive investment opportunity from the OTC market. At first they were skeptical and wanted more information but I insisted they work off of the chart itself. Here are some analysis they had done based on the chart:
Basic DOW theory shows that the company is making higher high and higher lows, meaning the upward trend is still intact. There is no sign of a reversal according to DOW theory but that is sometimes hard to spot, especially without volume as a secondary indicator - Charlie
With this company i have no idea where its going. Interesting that the stock almost tripled in the last quarter and it could be an indication of more momentum coming or it could just be overvalued (is the company really worth 3 times as much as it was 3 months ago??). - Kyle
I would wait for a better price, and perhaps a clear entry signal from the candles. In my case I would look for a bullish engulfing in the near future or maybe a pin bar [...] Keep in mind that it does show a bearish shooting star and engulfing in the recent past - Charlie
It seems like my friends have quite a lot to say about the chart when it's randomly generated, and who doesn't? This aspect of random price movements is one of the main criticisms of technical analysis. Many advocates often refer to it as an art rather than a strict, rigorous tool with a strict set of answers. I'm a huge fan of technical analysis and I constantly use it in my forex trading as I believe price patterns do indeed converge on market data. However, if we analysts can see patterns in both real market data and a randomly spun up chart, shouldn't we approach our trades with a bit more skepticism?