The mystery of patterns

I love patterns. As a mathematician, it seems natural that I seek to find patterns. Does this make me unique? Not really, it's a human trait to want to find a modicum of order in anything. In the book The (Mis)Behaviour of Markets: A Fractal View of Risk, Ruin and Reward by the late mathematician Benoit Mandelbrot, he illustrates this point through the use of a simple example. He plots a number of graphs. Some are of real market prices and some are random paths. I've repeated the exercise below. Which of the following examples are random and which are real?

It's somewhat difficult to guess, although I'm sure some of you might recognise some of these. At the bottom of the article, I've given the identify of the random plots, but I'll let the real ones remain a mystery (e-mail me if you really want to know their identities!). 

So is it futile to try to find patterns in markets? Well, no (I'm hardly going to say yes, when I've spent the past decade doing exactly that!). The question is not necessarily whether we can find patterns in price action, but understanding whether there is a logical rationale for patterns you find. There are numerous examples of patterns I can cite. Whilst some might be sceptical of technical analysis, there is an obvious reason why it works, the element of self-fulfilment. If enough market participants think a certain type of price action could result in a trend, and they jump on that trade, it becomes like a virtuous circle.

More broadly, there can be relatively intuitive reasons why we see certain patterns in price action, related to more behaviour aspects of trading. One interesting example can be seen in intraday volatility in FX markets. We could of course also make the case that volatility as a quantity is easier to forecast than market prices, because it has many "nice" properties. Volatility for example tends to be mean-reverting and also (relatively) bounded. In the plot below, we calculate the intraday volatility by time of day in EUR/USD over the past decade. There are several obvious patterns. Volatility tends to be lower during Asian hours, when there are generally fewer market participants. During London hours, volatility picks up and starts to tail off when London traders go home. We also notice spikes such as 1.30pm and 3pm LDN time, which tend to be the time of US data releases. Of course, attempting to monetise this pattern is tricky! However, it does illustrate that the idea that markets can have fairly distinct patterns.

So just because patterns might be difficult to separate from randomness, we can't simply say they don't exist in markets! There is of course the crucial caveat, that on many occasions there really is no pattern to be found. It is experience, which helps separate the data mining from the pattern finding in markets. Best of luck finding patterns in markets!

If you'd like to know more about vol patterns in FX, I wrote an article in the Financial Times explaining the volatility chart in a bit more detail (FT: Unpicking higher currency volatility: a guest chart - 10 Apr 2015) and also a comprehensive Thalesians quant paper discussing various patterns in both intraday currency volatility and also liquidity, which is very topical at this time (Thalesians: Once upon an intraday - 09 Apr 2015). If you'd like a copy of that research paper, let me know!

Like my writing? Have a look at my book Trading Thalesians - What the ancient world can teach us about trading today is on Palgrave Macmillan. You can order the book on Amazon. Drop me a message if you're interesting in me writing something for you or creating a systematic trading strategy for you! Please also come to our regular finance talks in London, New York and Budapest - join our Meetup.com group for more details here (Thalesians calendar below)

17 Apr - Budapest - Impact of bitcoin - Tamas Blummer & Panel featuring Izabella Kaminska / FT)
22 Apr - New York - How Smart Money Invests and Market Prices Are Determined - Lasse Pedersen
29 Apr - London  - Global macro & UK election panel - Eric Burroughs / Reuters, Mark Cudmore / Bloomberg, Jordan Rochester / Nomura, Jeremy Wilkinson-Smith / Independent & Saeed Amen / Thalesians

The random plot is graph 4. All the others are of real markets.