Event Time: Freeing Market Charts From the Time Clock
It looks like a normal chart, but there's a difference.
Almost all charts are denominated in time. We can have price on the Y-axis, we can show an oscillator or indicator on the Y-axis, but the X-axis invariably reflects time.
Time is a chronological event, whether we measure time in birthdays or in the behavior of cesium atoms.
Where is it written in stone that markets move in chronological time? Lopez de Prado explains that our reliance on chronological time is "rather arbitrary", as it reflects the role of the sun in agricultural societies. He describes the "new paradigm" in market analysis as one in which we move from chronological time to event time.
Indeed, there have been efforts to liberate market charts from the strictures of chronology. Point-and-figure charts draw a new bar when a threshold market move is made, making market movement the X-axis. Richard Arms adjusted the width of chronological bars for their volume, thereby changing the scale of the X-axis. Lopez de Prado's high-frequency trading directly denominates time in volume, as each X number of contracts or shares represents a time unit. As he points out, this brings a number of statistical advantages, including eliminating intra-session seasonal effects (e.g., changes in volume and volatility as a function of time of day).
Once we liberate the X-axis from chronological time, we open ourselves to the graphical display of many relationships. Just as the X-axis can bucket movement or volume, it could reflect units of sentiment change, volatility, correlation change, relative strength change, etc. My experience is that such event time charting displays relationships that are not immediately apparent when looking at standard charts based on chronology.
So back to the above chart of the last week in ES. A fresh bar is drawn every time the price of ES moves 500 times. During busy periods, we see more bars; during slow periods, we see fewer. If you had a trading system that went long or short with an X-bar signal, the system would give you more trades in busy markets--and busy market times--and fewer trades during slow markets.
In other words, the chart would not only normalize market action statistically, but it would normalize the trader's behavior, creating fewer trades when markets offer less movement. That makes the event-time chart a psychological tool, as well as an analytical one.
I will be posting more on event-time analysis in the near future.
Further Reading: Overtrading and Market Expectations
.
Comments
Post a Comment