Article Details
  • Published Online:
    July  2025
  • Product Name:
    The IUP Journal of Accounting Research & Audit Practices
  • Product Type:
    Article
  • Product Code:
    IJARAP080725
  • DOI:
    10.71329/IUPJARAP/2025.24.3.152-168
  • Author Name:
    Kaival Dave and Ruchita Patel
  • Availability:
    YES
  • Subject/Domain:
    Finance
  • Download Format:
    PDF
  • Pages:
    152-168
Volume 24, Issue 3, July-September 2025
Improving Trading Profitability of Option Strategies Based on Stock Price Prediction Using ARIMA Model
Abstract

Price prediction in security market depends on a lot of factors like risk-return relationship, position sizing, investment requirement, etc. This paper implements option strategies based on price predicted by ARIMA model to test its performance. It predicts the prices of Nifty 50 index as representative of fifty large-cap companies based on the ARIMA model. The best ARIMA model was selected based on Python code of Statsmodel. The model was tested with mean absolute percentage error (MAPE) and root mean squared error (RMSE). The option strategy was implemented based on price predicted by the ARIMA model. The performance of the strategies was compared with blanket selection of strike price without ARIMA. The strategy outcomes were then compared based on descriptive analysis, binomial test, Sharpe ratio and Sortino ratio. While the strike price selected based on ARIMA model provides more return compared to other strategies, strike price selection without ARIMA scores lower standard deviation. The win rate of all strategies was lower, but ARIMA modelbased strategy performed better due to higher profits in fewer trade. The study provides valuable insights into option strategies based on price prediction for traders and helps them improvise their overall return from the trade.

Introduction

Stock price prediction is crucial for any security market. Security market is considered as the easiest way to earn money in the hardest way. Irrespective of deployment of methods, prediction always comes from a study of historical patterns. Time series analysis is a broader concept used for predictive pricing. Establishing a mathematical model as a deterministic model based on physical phenomena is very much possible (George, 2015) .