Case Study

Designing a solution to predict the range of Bank Nifty

The Client

A broking firm engaged in derivatives trading.

The Task

The client was into derivatives trading and wanted to develop a statistical model which could help him predict the movements of the index in the day basis the highs or lows in the first hour. This would help the client take a favorable position in order to maximize his gain.
The client also wanted a dashboard to be prepared which would help him analyze the trends and the data.

The Approach

WWe extracted last 5 years of Nifty Data in order to ensure that at least 1 economic cycle was fully covered and that we had sufficient data points to run a detailed analysis.

We took a two step approach towards this:

Step 1- Testing the null hypothesis that the movement of index in the first hour had no correlation with the range in the subsequent hours.

Step 2- If the null hypothesis was false, then develop a scored model to predict the range of the index in the subsequent 2 hours of trading.

The Solution

A predictive model developed using SAS along with a simple to use dashboard
We used SAS as a tool and used log regression technique to come out with the model which could predict the range of the index on the basis of the highs or lows achieved in the first trading hour.
The probabilities were then scored in order to simplify the usage of the solution. A custom dashboard was also prepared to the user to provide the information is a user friendly way.

The Impact

  • The client was able to better plan his positions around trading

Let's talk about your queries.