#182: Random Runs

S2N Spotlight

Randomness has fascinated me for decades. I would argue this is one of the most complex areas of life, not just in quantitative finance. What makes it so tricky is that not only is it difficult to identify mathematically but also intuitively.

I have yet to meet a trader or investor who doesn’t think that the markets have noticeable patterns. It is the lifeblood of the industry, yet it is very difficult to get consistent and predictable answers when measuring randomness. While I believe that there are patterns. I would suggest that it is not nearly as simple as people think—a head and shoulders here and a trend there.

Today I am applying the Runs Test as the method for establishing randomness. I am not suggesting for a minute that this is the only test to apply but it is certainly one of the classics. This is a bit technical and probably not everyone’s cup of tea. I am currently working hard on a particular trading system so thought why not talk about what I am thinking about.

The Runs Test is a statistical method used to check whether a sequence of data is random. A run is a consecutive series of similar values, such as a string of positive returns followed by a string of negative returns. In financial markets, the Runs Test can determine if daily returns are random or show patterns. For example, if the sequence of “up days” and “down days” alternates too much or clumps together in an unusual way, it might suggest that the returns are not random.

The test compares the observed number of runs in the data to the expected number of runs if the data were random. It calculates a Z-score to measure how far the observed runs deviate from randomness, and a p-value determines if the deviation is statistically significant. If the p-value is less than 0.05, the test concludes the sequence is not random. Conversely, if the p-value is 0.05 or greater, the sequence is considered random, meaning it behaves unpredictably, as you might expect from random events like coin tosses.

I applied the Runs Test to a number of different assets, each symbol produced unexpected results. The takeaway from all of this is something I have preached for a long time.

Randomness is everywhere, however predictable patterns do exist but here is the bitter pill to swallow. Predictable patterns are marginal and very hard to find. Sorry I have to run 😎

S2N Observations

Will spare you today but take note how the commodity futures rallied yesterday.

S2N Screener Alerts

The Indian Nifty 50 is down 7 days in a row. It has only happened 57 times in 29 years. This is considered a Run in a Runs Test, just saying.

Performance Review

For those who are new to the letter, the shading is Z-Score adjusted so that only moves bigger than usual for the symbol are highlighted.

News Today