The leading knowledge platform for the financial technology industry
The leading knowledge platform for the financial technology industry

A-Team Insight Blogs

Vulpes Investment Management Revamps IT for Systematic Trading

Vulpes Investment Management is revamping its technology infrastructure as it moves out of volatility trading and into systematic trading. The revamp includes implementation of OneMarketData’s OneTick stream processing and tick data management solution for quantitative research, modelling and strategy development, as well as a Tora electronic trading platform and portfolio management system, Redis in-memory database technology and use of the Python programming language.

The firm is also switching its market data vendor for the systematic fund, ending an arrangement with Bloomberg, which it found inflexible in terms of licensing for a small operation, and considering other providers such as Interactive Data, Thomson Reuters, S&P Capital IQ and Morningstar. It expects to decide on a market data vendor in the next couple of weeks.

The systematic trading fund, which has yet to be named, is based in Singapore and is the only high end, quant based trading fund among five funds offered to investors by Vulpes. It trades intraday dislocations and mis-pricings on Asia Pacific exchanges, monitoring prices, volumes and spreads in real time and predicting movements over the next one to three hours, usually closing positions out by the end of the day. Its investors are predominantly high net worth individuals in Singapore and it will initially trade cash equity and futures securities on exchanges in Tokyo, Sydney and Singapore. It plans to add exchanges in Asia Pacific, although any additions must prove beneficial in terms of alpha generation versus trading costs, and in the longer term may extend the asset base it trades and market the fund outside Asia.

Back to today, and Vulpes is testing trade flow through its systematic trading system ahead of full production planned for January 2014. Scott Treloar, chief risk officer at Vulpes and an experienced quant, is one of five employees working on the systematic trading fund out of a total of 15 at the firm. He explains: “The firm was set up in 2010 with a focus on volatility trading, but volatility has been crushed by banks and is on a downward trend. So we are moving out of volatility and into systematic trading, but sticking with the Asia Pacific market as we have a good understanding of it and it is not very liquid, which keeps the big guys out.”

Treloar joined Vulpes a year ago to lead the systematic trading group and develop a suitable trading system. He says: “Historically, we were a point and click trading group, but we are moving to an infrastructure that provides straight through processing and a real-time view of our positions and risks. We will touch nothing.”

Considering the large volume and complexity of tick data needed to support systematic trading strategies, Treloar decided to move away from the HDF5 database associated with Python and select a specialist database provider. The firm considered Kx Systems’ kdb+ database, SciDB, a database designed for time series data and scientific research, and OneMarketData’s OneTick. It ruled out kdb+ because it would require the team to learn a new database language and SciDB on the grounds that it is a new system requiring a little more development.

On this basis, OneTick was chosen for its performance and a Python application programming interface (API) that supports integration with the fund’s Python framework. Treloar says: “With OneTick, we can conduct research, find opportunities for alpha and test our strategies across multiple asset classes efficiently and using one solution.”

Detailing how OneTick is configured to work for the fund, he says: “We stream the tick data into our OneTick in-memory database and, overnight, the in-memory data is loaded into the archive database. We query the in-memory data using the OneTick event processors to aggregate, filter and transform the streaming data quickly for use in our proprietary trading algorithms. We update and recalibrate the algorithms’ parameters overnight with historical tick data and we use up to one year of level one tick data in the development and cross-validation of our trading strategies.”

The fund develops its trading algorithms using the open source Python programming language and has built position sizing and risk management, which are integrated with trading signal generation. It also uses Python to link the components of its trading and risk infrastructure, the idea being to create an open architecture of replaceable components with a single programming language linking them together.

Another open source element of the system is a Redis in-memory database. This was chosen for messaging above a solution from RabbitMQ (a unit of Pivotal) and is used to control and manage positions and orders. Orders generated by the system are fed through an API into a Tora Compass electronic trading platform and, once executed, trades flow into Tora’s Prism portfolio and risk management system that allows positions and risk to be viewed in real time. The team has developed a web application that pulls data out of Prism and slices and dices it so that it can be seen in many ways, but it is considering adding a data visualisation tool such as Panopticon or Tableau to display the data.

In terms of latency, Treloar says the fund operates in real time, but will not play in the latency game. He explains: “While we are not ultra-high frequency, we do trade intraday, so it is important to be able to operate in ‘event time’. Our trading strategies are processing events or ticks and necessitate making trading decisions in real time. This means our infrastructure, from market data management to execution and reconciliation, needs to be highly automated and integrated to allow the monitoring of positions and risks in real time.”

While the systematic trading fund is a new departure for Vulpes, its use of OneTick is a stepping stone for OneMarketData, which already has one quant fund customer in Singapore and supports Standard Chartered Bank in foreign exchange market making and electronic trading in Singapore, London and New York. Richard Chmiel, senior vice president at OneMarketData, says: “We are beginning to gain critical mass in Asia, which is a big target market for us, but at the moment generates a small amount of revenue. The prospects for Asia are good as the appetite for risk is growing and there is more investment in technology and people.” OneMarkeData has people on the ground in Hong Kong, Tokyo and Melbourne, and is planning to recruit more as well as send employees to Asia on a temporary basis in January to support OneTick proof of concepts it is working on with prospects.

Beyond Asia, Chmiel says 2013 has been a good year for OneMarketData in the U.S. and Europe, and he expects the company to have equal or better success in 2014, with revenue split between the sell side and buy side. Among plans on the drawing board that could contribute to customer gains and revenue growth are hosting OneTick and developing additional components for the software.

In terms of hosting, OneMarketData does not plan to host OneTick itself and will instead look for hosting partners. To date, it has an arrangement with Lime Brokerage that hosts OneTick for U.S. equities and it is considering similar relationships in Asia and Europe that could come to fruition in 2014.

Similarly, the company’s product plans are in a conceptual phase, but if development is successful new products could move into production next year. Chmiel says the company is working on a reference data solution that would build on OneQuantData and could be hosted with OneTick. Considering the activities of a typical quant fund – namely research, model development, risk management and order execution – he notes that OneTick covers research and model development, but could, perhaps, be developed to fulfil the needs of risk management and order generation.

While OneTick has risk management capabilities, Chmiel says most users handle risk with a different system. He explains: “If we developed more processors for risk management we could make OneTick risk management seamless and provide an out-of-the-box solution.” Additional components could also be added to OneTick to generate orders and send them to market gateways, consolidating the trading process into fewer systems. Combining OneTick and these extensions with the possibility of hosting, Chmiel says: “Small firms have an appetite for solutions like this as they reduce both time to market and cost.” What he doesn’t say is if and when OneMarketData will deliver such solutions.

Related content

WEBINAR

Upcoming Webinar: Market data management, licensing and administration in the post-Covid environment

Date: 25 May 2021 Time: 10:00am ET / 3:00pm London / 4:00pm CET Duration: 50 minutes Market data administration has always been a challenge. For many firms, keeping tabs on permissioning and entitlements, compliance with licensing agreements, and reconciling all that with increasingly complex invoices requires a significant dedicated resource with a clear understanding of...

BLOG

A-Team Webinar to Examine Trade Surveillance in Today’s New Normal

A panel of trading surveillance experts will discuss lessons learned and best practices for ensuring compliance in work from home (WFH) environments on A-Team Group’s April 8 webinar: Trade surveillance: Deploying monitoring and surveillance capabilities for today’s new normal. The webinar – sponsored by OneMarketData and ACA Compliance – will discuss the trade surveillance and...

EVENT

Data Management Summit London

The Data Management Summit Virtual explores how financial institutions are shifting from defensive to offensive data management strategies, to improve operational efficiency and revenue enhancing opportunities. We’ll be putting the business lens on data and deep diving into the data management capabilities needed to deliver on business outcomes.

GUIDE

Entity Data Management Handbook – Seventh Edition

Sourcing entity data and ensuring efficient and effective entity data management is a challenge for many financial institutions as volumes of data rise, more regulations require entity data in reporting, and the fight again financial crime is escalated by bad actors using increasingly sophisticated techniques to attack processes and systems. That said, based on best...