Estratégias de trading baseadas em médias móveis: uma análise exploratória com PETR4

Abstract

The present work seeks to implement and validate some trading strategies based on moving averages with PETR4. The highest frequency series is minute by minute and the period under review is from 12/2010 to 10/2018, from intraday data collected from the Bloomberg. It was used trading strategies based on only one moving aver-age; based on two moving averages, one to purchase and another to sale; and the use of indicators developed by Bill Williams, fractal and alligator. The implementa-tion was validated computationally, through an algorithm developed and presented on the work. As a result, there were negative returns for the highest frequencies (1min, 3min and 5min); the lowest frequencies (60min and 120min) reached the highest number of positive returns; the best performance seen was 2, i.e., 200% over the 95-month period, i.e. 7 years and 11 months, in the strategy using only one mov-ing average of size 8 and frequency of 60 minutes. Finally, regarding the sensitivity of the return to the days of the week, there was no significant distortion; however, for the daily hours, there are less favorable returns at the beginning of the day.