Langsung ke konten utama


Menampilkan postingan dari Desember, 2019

Applying Corrective AI to Daily Seasonal Forex Trading

  By Sergei Belov, Ernest Chan, Nahid Jetha, and Akshay Nautiyal     ABSTRACT We applied Corrective AI (Chan, 2022) to a trading model that takes advantage of the intraday seasonality of forex returns. Breedon and Ranaldo (2012)   observed that foreign currencies depreciate vs. the US dollar during their local working hours and appreciate during the local working hours of the US dollar. We first backtested the results of Breedon and Ranaldo on recent EURUSD data from September 2021 to January 2023 and then applied Corrective AI to this trading strategy to achieve a significant increase in performance. Breedon and Ranaldo (2012) described a trading strategy that shorted EURUSD during European working hours (3 AM ET to 9 AM ET, where ET denotes the local time in New York, accounting for daylight savings) and bought EURUSD during US working hours (11 AM ET to 3 PM ET). The rationale is that large-scale institutional buying of the US dollar takes place during European working hours to pa

US nonfarm employment prediction using RIWI Corp. alternative data

Introduction The monthly US nonfarm payroll (NFP) announcement by the United States Bureau of Labor Statistics (BLS) is one of the most closely watched economic indicators, for economists and investors alike. (When I was teaching a class at a well-known proprietary trading firm, the traders suddenly ran out of the classroom to their desks on a Friday morning just before 8:30am EST.) Naturally, there were many efforts in the past trying to predict this number, ranging from using other macroeconomic indicators such as credit spreads to using Twitter sentiment as predictive features. In this article, I will report on research conducted by Radu Ciobanu and I using the unique and proprietary continuous survey data provided by RIWI Corp. to predict this important number. RIWI is an alternative data provider that conducts online surveys and risk measurement monitoring in all countries of the world anonymously, without collecting any personally identifiable information or providing incentive

Experiments with GANs for Simulating Returns (Guest post)

By  Akshay Nautiyal, Quantinsti Simulating returns using either the traditional closed-form equations or probabilistic models like Monte Carlo has been the standard practice to match them against empirical observations from stock, bond and other financial time-series data. (See Chan and Ng, 2017 and Lopez de Prado, 2018 .)   Some of the stylised facts of return distributions are as follows: The tails of an empirical return distribution are always thick, indicating lucky gains and enormous losses are more probable than a Gaussian distribution would suggest.  Empirical distributions of assets show sharp peaks which traditional models are often not able to gauge.  To generate simulated return distributions that are faithful to their empirical counterpart, I tried my hand on various kinds of Generative Adversarial Networks, a very specialised Neural Network to learn the features of a stationary series we’ll describe later. The GAN architectures used here are a direct descendant of the