Fingerprint of a Crash: Recurrence Plots

Recurrence plots provide a unique way to visualize time-series data. In the right graph, we have the price of 3D Systems (DDD) from 2007 to now. On the left we have a recurrence plot (RP) of DDD’s price. RP’s are similar to a correlation heatmap:

  • Each axis is time, the same time range as the price chart on the right
  • The color of each point represents how similar the x time is to the y time in terms of price movement
  • The plot is symmetric around a diagonal; Each time is most similar to itself by definition

The darker the color the more similar the timesUntitledWe can use a recurrence plot (RP) to visualize the effects of crashes over time. Notice the cross pattern where x = 2011 and y = 2011? That is the fingerprint of the crash in DDD’s price in early 2011. There are crosses in 2013 and 2014 as well which correspond to severe declines during the same time periods.

The cross for each individual crash intersects those of other crashes. Notice how dark the patches are around the points ( x = 2011, y = 2013 ) and (x = 2011, y = 2014)? The RP is telling us that these periods were similar. Visual inspection of the stock chart can verify their similarity.

The first cross in the RP took about 4 years to occur. The next occurred 2 years later, then another in one year. The time between successive boom-bust cycles is shortening. We could be approaching what Didier Sornette would call a “phase-change.” This is a point where the price process fundamentally changes. You might also call it a crash.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s