Discover how to build a trend following strategy in the SigTech platform.
Watch a step by step demo as SigTech Product Manager, Estèphe Corlin, demonstrates how to build an investment strategy trading 52 highly liquid futures applying a trend following signal across the asset classes equities, fixed income, FX and commodities.
You can review the strategy in our interactive performance tool and read the full video transcript at the bottom of the page.
Video transcript
Using SigTech’s quant technology platform, we will walk through the construction and backtest of a bespoke trend following strategy.
Our investable universe is defined by a set of 52 highly liquid futures spanning the asset classes equities, fixed income, FX and commodities. SigTech provides users with default strategy objects, which are pre-built systematic investment strategies and functionalities. One example of this is es_index_front, which maintains constant exposure to the S&P 500 futures strip, by continuously rolling into the front future. The plot below shows the performance of this default strategy over time.
A set of default rolling future strategies are defined as our starting universe. The universe is then supplemented with another set of futures groups, for which the strategies are built via the building block RollingFutureStrategy and RollingFutureFxHedged. These objects handle the rolling of one futures contract to the next over the backtesting period, in a customizable and transparent manner.
SigTech provides ready to use functionality for multiple trend indicators, such as RSI, MACD and different filters like L1 and Hodrick-Prescott. For this strategy, we will compute a custom trend indicator that calculates the average direction of a time series based on a 60 day trading window using the following formula:
The number of daily upwards returns minus the number of daily downward returns divided by the 60 day rolling period.
Our interactive signal report provides different analyses for quickly investigating the potential and effectiveness of our trend signal. The quantiles tab can be used to showcase how well the signal predicts the next day’s directional move. The signal values are ranked into quantiles, and the upward sloping line shows that the stronger the trend signal, the more likely it is to continue over the next period. We can thus conclude that – at least in-sample – our trend signal can be used to generate alpha.
To construct our portfolio, the signal values are first scaled based on extreme observations over a rolling three year window, resulting in an indicator showing values between 1 and –1. We then apply an inverse volatility weighting, scaling the signal to weight the assets inversely to their risk, targeting 5% overall portfolio volatility. Our volatility measure is defined as a rolling 60 day exponentially weighted standard deviation.
To account for asset correlations, the weights are scaled once more by targeting an overall volatility of 5% based on a 60 day covariance estimate between instruments.
Finally, our backtest is run using the signal strategy building block. A 2% threshold is set to trigger a rebalancing, and a bid-ask spread of 5bps is applied to account for transaction costs. The strategy achieves a 4.5% annualized return with a 0.9 sharpe ratio. More notably, the yearly returns exhibit only four down years over the entire 20-year backtesting period.
Get in touch to find out more about how you can build bespoke trend following strategies on the SigTech platform.
This document is not, and should not be construed as financial advice or an invitation to purchase financial products. It is provided for information purposes only and is subject to the terms and conditions of our disclaimer which can be accessed at: https://sigtech.com/legal/general-disclaimer
