Intraday data at SigTech
Data management is at the core of SigTech’s quant technologies platform; providing access to over 250 research-ready intraday time series across all liquid asset classes. The data is available down to 1-minute bars and is easily resampled into lower frequency intervals.
SigTech facilitates the assessment of intraday data and in some instances, its adjustment. The platform enables users to select from a library of scoring methods (e.g. percentage moves, Z-scores, stale value) or to define their own, when filtering out low credibility data points. Thus, users can decide between using the raw data as supplied by the exchanges or, data filtered and adjusted according to SigTech’s or their own specifications.
SigTech’s intraday offering: data coverage
SigTech’s intraday offering: asset class coverage
Accelerate your research process
SigTech’s quant technologies platform is structured to follow a traditional investment process, from data management to backtesting and production. Core functionality combines with a wide range of customisable strategy templates to accelerate the investment process and reduce time-to-market.
Adopting a breakout methodology, we now illustrate the development of a cross asset trend following strategy on the SigTech platform
Intraday Breakout Strategy
The breakout strategy is implemented using intraday data. It follows the rationale that a trend at the open (positive or negative) will persist across the entire trading day.
The strategy makes use of a breakout rule based on today’s opening price compared to yesterday’s closing price:
If open > previous close (1 + z std), open a long position and close it at the end of the trading day
If open < previous close (1 - z std), open a short position and close it at the end of the trading day
Otherwise take no action
open = open price for an asset on day T
previous_close = close price for an asset on day T-1
z = model parameter
std = standard deviation of an asset over the previous 20 trading days
Once a condition is met, the strategy executes the long or short position with a time-lag of 1 minute and closes it at the official market close on the same trading day.
In our tests we set the z value to 2.2. The strategy will:
initiate a long position if the open price is more than 2.2 standard deviations above the previous close
initiate a short position if it is more than 2.2 standard deviations below the previous closing price.
The z value of 2.2 is selected after conducting a comprehensive sensitivity and robustness test.
During backtesting we apply a transaction cost model taking commission and market impact estimations into account.
Performance overview - single asset strategy
We first test the breakout strategy for a single asset (S&P 500) and also address the strategy’s sensitivity to changes in the z value.
The strategy using only S&P 500 futures is backtested over the previous 22 years and the output is summarized in the graph and table below. Running a single asset strategy fails to satisfy the demands of a well-diversified, systematic investment strategy. However, its Sharpe ratio of 0.82 is noteworthy.
Performance overview of S&P 500 strategy
Risk/-return overview of S&P 500 futures strategy
To assess the strategy’s sensitivity to changes in the z value, we run a sensitivity and robustness analysis. The graph below demonstrates a positive correlation between the z value and the Sharpe ratio; the more extreme the jump (positive or negative) at the open, the larger the intraday trend. The reason for not using a higher z value - which, according to the sensitivity analysis exhibits a higher Sharpe ratio - is that higher z values correlate with fewer trades. Systematic investment strategies seek to capture profitable patterns multiple times, rather than relying on a few observations.
Sensitivity of Sharpe ratio to the z value
Performance overview - basket strategy
Our second step creates a basket strategy using 17 futures across the asset classes equities, fixed income, and FX. To determine portfolio weights for each asset we apply an inverse volatility allocation method.
It is worth noting that the z value is the same across all markets. As shown above, we performed robustness tests across all assets to determine the model parameter, i.e. the sensitivity of results to changing z values.
The graph and tables below show the basket strategy’s performance since 2013. The strategy accounts for direct and indirect transaction costs and exhibits a Sharpe ratio of 0.91. It posted a positive return in 6 out of 11 years.
Performance basket strategy
Risk/-return overview of basket strategy
Return overview of basket strategy
Once the research phase is complete, investment managers often face a long time–to-market due to a fragmented internal system landscape. SigTech significantly reduces this lead time to launch a new strategy by enabling one-click strategy deployment (essentially using a single line of code). This eliminates the expensive and time-consuming interplay between software engineers, devops engineers and researchers. In addition, an integrated platform for research and production reduces the disconnect between investment results achieved in the research and the live production environment, respectively.
SigTech’s quant technologies platform accelerates and streamlines the research process by managing the data requirements of investors and providing 100+ strategy templates and a wide selection of research functionalities across all liquid asset classes and a wide range of financial instruments. Examples of common investment strategy use cases our clients are implementing are global macro, tail risk and custom indexing strategies. Our code is open source and fully customisable to the individual specifications of the user. The SigTech platform frees investors to focus on the core activities of research and alpha generation.
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://www.sigtech.com/legal/general-disclaimer
Key features used within SigTech’s quant technology platform
To learn more about the templates and functionalities used to create and test our strategy, please follow the links below.
One of the key functionalities of SigTech’s platform is the easy creation of customized, historical rolling futures time series, both for end-of-day and intraday data. We use the IntradaySignalStrategy template that has two key input variables:
Signal object - function using the local currency rolling future template
Tradable instrument - strategy using the fx hedged rolling future template
To accurately implement the signal rule at the exact trading times (open and close, respectively) throughout the backtesting period, our platform provides time variant session data for futures via a dedicated API method.
Strategy cache service
To analyze the signal (at asset and basket strategy level) it is common practice to run multiple simulations. It is therefore essential to have the strategy template and market data available without having to re-run a backtest for each new alteration. This is enabled by using the strategy cache services which stores the strategy objects (including market data) beyond the lifetime of the current research session. This significantly accelerates the research process.
Intraday signal strategy
The intraday signal strategy enables users to specify the exact time an investment decision should be taken and to execute an order at the closest possible intraday time, i.e. during the next minute. The instruction is easily altered using the template’s input parameters.
The performance report generates a comprehensive and fully customisable breakdown of a strategy’s risk and return metrics.