Finding 'edge' through Data Curation

When we talk of ‘edge’ in trading it essentially means, what is it different that you or your system has which would lead to an ‘alpha’ in terms of returns. One way to extend that question is to ask ourselves, as to what are the sources of that edge?

To me having a more deeper and nuanced understanding of the contexts and setups that I trade, exponentially adds to my edge. To get a better sense of my contexts, one of the practices that has massively helped me in my trading, is curating setup specific data. This is a lot of work, let me tell you. Sometimes it’s very difficult to train the computer to do what we humans can do intuitively, which means a lot of it is manual labor.

Let me try to give you a sense of what I mean, if you follow my blog you would know that, these are the three setups that I trade.

NIFTY - Scalping Set-up - 01 - Opening Spikes & Opening Drive

NIFTY - Scalping Set-up - 02 - Mid-Day Mean Reversion

NIFTY - Scalping Set-up - 03 - Afternoon Range Extension

Now each of these setups have their nuances and details, like

a) At what time did the entry get signaled? Is that time range bound? Is there a seasonal skew to it?

b) Range breakouts on VIX? Time and Amplitude.

c) What is the average size of the pullbacks that in the setups you trade?

d) What is the ideal holding time for your some of your setups, based on the length of the trends?

A lot of these computations are possible only if you have specific data. Therefore, this is a practice that we follow in-house and for our clients - i.e. to capture such data so as to run tests on it.

Here is a mini snapshot of the data. (Disclaimer : The snapshot may make no sense whatsoever without context)

View fullsize

This practice when followed over long periods of time can give you a gold mine of data, which (I believe) can add to your trading expectancy.

Article Excerpt & NiftyScalper Data Analysis - Fundamentals of Short-Term Trading : Part Two | Dr. Brett N. Steenbarger

Ever since I began my journey in the world of trading, articles and books by Dr. Steenbarger has been of great value to me.

Coincidentally I came across this excerpt where he is alluding to the idea of looking at markets horizontally, something that one of my team members' tried analyzing sometime back. First things first, here is what Dr. Steenbarger has to say.  

Article link - http://www.brettsteenbarger.com/Short-Term%20Trading2.doc

All the seasoned day traders that I have interacted with, either intuitively or through analysis, have said the same the thing, ie Not all time slots are the same.

Lets look at the volatility spikes on an Intraday basis for NIFTY. This was 2 years data, and we tried to do the same horizontal analysis by normalizing each day's move. You would clearly see two peaks and a trough. Of course this does not give a nuanced picture as a Monte Carlo analysis would, but still tells us quite clearly where the "meat" is. 

On a closing note, I was once told by a developing trader that he take trades only after 10:15 am IST in NIFTY, as he "believes" that's when you get an opening range breakout. Needless to say, he would get stopped out more often than not, and perhaps would see his trade move in the intended direction, couple of hours later. 

A simple horizontal analysis can help you time and size your trades optimally. Try it out.