MarketLife Ep 41 – Randomness: why it matters and how to understand it

I received a good question from a listener, Alex:

I was talking to a friend of mine (who is a physicist). ” I presented him with some market statistics that seem to hold true in many markets, e.g. that the market has only a low probability to hit yesterday’s high and low on any given day (10-15%, depending on the market). My friend argued that in order for a statistic to be significant, I’d have to test it on randomly generated market data. Only if the random data does not show the same tendency, a statistic can be thought of as showing a bias. Do you think that my friend is right? Are statistics only significant if randomly generated data does not show the same tendency? What if both, real historical data and random data show the same tendency? I’d love to hear your thoughts on this. Best, Alex”

This podcast ended up being a bit more involved than most, but I hope you find it useful and entertaining!

Here is a link to the academic paper I mentioned in the podcast.

If you enjoy the podcast, one of the very best things you can do for me is to leave me a review on iTunes here. Also, if you like the music for this podcast, then be sure to check out Brian Ashley Jones, my friend, and a fantastic singer-songwriter.

Enjoy the show:



Adam Grimes has over two decades of experience in the industry as a trader, analyst and system developer. The author of a best-selling trading book, he has traded for his own account, for a top prop firm, and spent several years at the New York Mercantile Exchange. He focuses on the intersection of quantitative analysis and discretionary trading, and has a talent for teaching and helping traders find their own way in the market.

This Post Has 2 Comments

  1. Glenn Lucero

    This topic is always the most intriguing to me and thank you Adam for writing about it. I love talking about and thinking about randomness. However this topic always brings me down the rabbit hole so let’s jump in shall we? Let’s say for example that the market is 100% completely random and there is no edge at all. Now, we do know that you cannot make money in a random market correct? In financial markets we make money when price trends, pulse and then pullback and then pulse. Characteristics of a trend or price sloping upward or downward. Randomness can also trend and can have large trends and trends that would appear to have an edge. Take for example this magician, he flipped a coin and made the coin land heads 10 times in a row. That sounds like and edge doesn’t it? Here is the link to prove how he did it by the way, which has quite a bit to do with trading. Anyhow, here is the million dollar question. If randomness can trend and trend significantly then can we time our bets and make money in the random market? Think about the guy at the craps table who just rolled for 45 minutes or ( actually happened to me, I won 23 hands in a row at a big limit poker game) It would appear so because when we have an edge we make the most out of that edge when the market is lucky enough to help us( trend most likely ) Correct? So the question is: Is it possible to make money in random markets if your timing is right?

  2. Bing Garcia

    Using daily stock return data from Jan 1, 1926 to Dec 31, 2014 the autocorrelation coefficient, p, is plotted for both value and equal weighted indexes of all U.S. stocks over 750 day rolling windows, roughly three year periods. The results show that stock market predictability waxes and wanes, instead of declining over time as the Efficients Markets Hypothesis suggests. The 1930s were a period of very little predictability, and p stayed inside the range of statistical insignificance (the interval around zero marked by the two dotted lines), so the Random Walk Hypothesis was a reasonable approximation of reality. Starting in the 1940s, p began to climb outside of that range, becoming statistical significant, and reaching a peak in the mid-1970s, after which it declined progressively, falling back into the range of statistical insignificance until the recent financial crisis. After the crisis, not only did p continue its decline, but it turned significantly negative. This implies day to day reversals, positive returns today predicting negative returns tomorrow. The market dynamic had changed. Adaptive Markets by Andrew Lo

Comments are closed.