The Gateway to Algorithmic and Automated Trading

Open Platform

Published in Automated Trader Magazine Issue 07 October 2007

When a US options prop shop built a new fast-track strategy development platform, its business model changed overnight. Tom Heffernan, Director, Global Marketing, Last Atlantis Capital Management, explains why the alternative asset manager encourages emerging trading firms to leverage its technology to perfect their own models.

Tom Heffernen

How has Last Atlantis's business model evolved from inception to the present day?

Last Atlantis originally started as a proprietary options trading shop back in 1999. We were primarily involved in high-speed arbitrage trading using US exchange-traded equity options.

As our primary playground was the US equity options markets, we experienced a rapid succession of changes at the exchange-level, including the introduction of new technologies, operations and pricing mechanisms. As a result, the strategies that had been profitable when we started eventually lost much of their edge. So to survive in this rapidly changing marketplace, we set out to develop a very robust technology infrastructure which would allow us to shorten the timeline for strategy and model development to accommodate changing conditions and improve our capture capability of opportunities through higher-speed data scanning.

How did this change your business model?

Once we built these systems, we realised we had developed a formidable, highly scalable infrastructure that could be applied to more than our proprietary business. We also recognised that the rapidly changing nature of the equity options markets may force - as it did - many other proprietary options trading firms to alter their business models. So we chose to diversify our business by leveraging new partnerships and our existing resources to a develop new revenue stream. With the technology and strategy development team from our prop business, we teamed up with Stig Ostgaard, an original Richard Dennis 'turtle', and Irwin Berger, both formerly of Sjo, to launch our alternative asset management firm, Last Atlantis Capital Management (LACM).

"… we're driven by opportunities presented from both sides: small funds and independent traders with potentially good ideas; and institutional investors who are seeking specific investments…"

Through LACM, we've developed a master feeder fund structure which allows us to introduce new strategies within our primary fund without any cross-liability between the individual strategies. The concept uses a Delaware Series LLC for our on-shore feeder and a BVI Segregated Portfolio Company (SPC) as its off-shore counterpart. The structure provides the flexibility critical to our mission. It allows us to better leverage the investment in our development platform by making it available to emerging traders with good ideas but without the infrastructure to test and validate them.

How did the firm's automated trading capabilities first develop?

Not all of our trading is automated; our proprietary trading uses automated systems in the form of high-speed scanning. Initially, we were seeking arbitrage opportunities in the US equity options markets that, to be traded successfully, required technology that would allow us to quickly identify opportunities. The exchanges (and their member firms) had issues handling the velocity of our trading and instituted rules which required manual execution of all option trades. So, based on the new rules, and to ensure we stayed in compliance with exchange regulations as they evolved (primarily to keep pace with systems developers), we engineered higher-speed scanning technology that would allow us to identify trades faster and still allow for enough time to capture and execute the trades manually. Within this operational and regulatory framework, we continue to take extra care as we develop or modify systems to ensure we remain in compliance with the applicable rules and regulations governing our business.

What role do automated trading strategies play in Last Atlantis's overall proposition to investors?

Automated strategies, of course, take advantage of our historical strengths, but they also provide a strategic counterweight to many of the discretionary programmes in our portfolio. Certain investors like to balance discretionary and non-discretionary strategies within their portfolios as different methods for sourcing alpha. As we saw in early August, many well-performing quant funds experienced a rough patch as their models initially didn't perform well within the more volatile environment.

New Ideas

How does Last Atlantis work with 'partner' funds, both in terms of model development and the proposition to investors?

It helps to understand that we're driven by opportunities presented from both sides: small funds and independent traders with potentially good ideas who need access to technology and a critical knowledge base to help them fine tune their programmes; and institutional investors who are seeking specific investments that fall within their investment mandate. Through our technology, personnel and marketing resources, we work to refine and subsequently marry the ideas with investment mandates specified by allocators.

What kind of automated trading models are you developing?

Because of the breadth of our investment programmes, it's a mixed bag. All of our trading models are fully customised and conditioned on market activity. Our trend-following models can incorporate volatility, momentum or inertia parameters. We use volatility arbitrage when market conditions are suitable. In other instances, we use a mix of models incorporating fundamentals and technicals.

"We didn't want to be dependent on the schedule or resources of an outside vendor to customise our technology so we could test a new idea."

There's not much alpha in off-the-shelf solutions, so we've spent tens of thousands of man-hours developing an ever-growing, proprietary suite of systems to further enhance performance of our trading ideas.

What are the ways in which your trading models (and discretionary funds) ensure risk diversification?

Depending on the investment strategy, we diversify risk any number of ways - by asset type, market, timeframe, fundamentals, market cap, volatility, baskets, money flow, technicals, strategy, liquidity, etc. Some systems incorporate currency hedges, and others use options as a hedge to directional plays.

To what extent is your technology infrastructure in-house built and why?

All of our platforms have been developed in-house. These platforms include our high-speed data networks, back-testing network cluster, analytics, scanners and trading front-ends. Because of our business model, we wanted to maintain control over our technologies so we could quickly adapt our platforms to analyse and test any type of strategy, concept or market. We didn't want to be dependent on the schedule or resources of an outside vendor to customise our technology so we could test a new idea. We have a specific plan of what we want do and we felt using external technology vendors for critical systems would further an already time-sensitive development cycle.

What backtesting techniques do you employ?

All of our backtesting techniques are custom-made and proprietary. We have built a highly robust historical backtesting platform that allows us to analyse systems across one or more markets within a comparatively short timeframe. This allows us to quickly recognise a system's value, including its strengths and weaknesses, so we spend less time watching data process and more time improving system performance.

What are the key challenges of trading on multiple markets?

The diversity of our programmes requires nearly complete global coverage of all exchange-traded assets. One of our daily challenges is maintaining data quality. In fact, there are few more important aspects, not only to building the systems themselves, but to trading them as well. Poor data can be very costly. We have built our own feed handlers to manage our raw data feeds. As such, we have built very sophisticated data filtering technologies that not only filter out bad ticks and ensure that our real-time and historical data systems reflect the real market, but also help us better understand individual market behaviour. We want to know if an anomaly is becoming normalised.

Before we begin trading a different market, we thoroughly research how it functions. Each market has its own volatility characteristics and you need to understand pricing characteristics (i.e. bid/ask spreads) to know where you really need to be to get executed. And because we have global coverage and an international client base, we carefully monitor currency risk. We've built risk systems which help identify and limit our exposure intra-day.

key challenges

What ways do investors/allocators select from your portfolio of fund products?

Initially, investors approach us to get an understanding of our platform. Most are used to dealing with shops with one or two strategies, and here we are currently sitting on 15. Many think we're a fund of funds; we're not, but we do have fund of funds offerings. Once we get past that, though, we start digging into their specific requirements.

Most of our clients have a very good understanding of their investment mandate. So they may be looking for a specific strategy or a non-correlated offering which may be unique or may fill a niche within their portfolio. As we dig further, many times we discover the client is looking for some type of structured or customised product which can be cultivated through a customised, multi-strategy product that incorporates two or more of our off-the-shelf strategies. For longer-term projects, we work with institutions in researching and developing proprietary strategies.

What trends do you see over the next 12 months and how will Last Capital respond?

Market integration and consolidation are forcing continued refinement and development of new investment programmes. You may see more systems move to a 24-hour trading cycle to better utilise capital and enhance returns. Also, we've seen US institutions exhibit greater willingness to consider emerging managers as a new source of alpha, as studies have shown that young programmes tend to out-perform in their adolescent years. This trend bodes well to shorten the payoff timeline to recoup the investment spent on incubating new automated systems.

"… studies have shown that young programmes tend to out-perform in their adolescent years."

Also, as we saw with the recent market turmoil, many boilerplate strategies got crushed en masse, showing that strategies which were once thought to be non-correlated are now in fact correlated, based on the number of funds using the same ideas. Continued development of new ideas will help investors better manage risk and keep their source of alpha. With our technical capabilities and business model focused on the continuous incubation and development of new systems, we can adapt to the on-coming era of disposable alpha.