Sun, Jan 20, 2019
A A A
Welcome Guest
Free Trial RSS
Get FREE trial access to our award winning publications
Horizons: Family Office & Investor Magazine

Company Portrait: Rotella Capital Management, a Leader in Quantitative Research

Sunday, March 11, 2018

Rotella Capital Management is a veteran systematic trading firm started in 1995 by Robert Rotella. Since then, the firm has been on the forefront of quantitative research - finding ways to sort through technology and use it to its greatest value. Rotella has operated both as a family office and a manager of outside money. As the firm looks toward the future, COO Ian Ram and Co-CIO Jagdeesh Prakasam say there is value in harnessing the new power of artificial intelligence for systematic investing, but the key will be identifying the best uses cases rather than expecting it to be a silver bullet.

 

The history.

Robert Rotella started his trading career in 1982 and began trading independently in 1985.

He began trading the Polaris program in 1990 with Commodities Corporation, one of the most successful incubators of traders in the industry. Louis Bacon, Paul Tudor Jones, and Bruce Kovner all started at Commodities Corporation. Now, more than two decades later, Polaris remains Rotella's flagship program with a 27-year track record of consistent performance.

Rotella places a high premium on research and pins the success of its trading programs on a research process that is in-depth and rigorous. "We have always been a research-oriented shop," says Ian Ram, Chief Operating Officer at Rotella Capital Management. "As a family office we invest a lot of our own money into our programs and we are constantly working on our research process to ensure that we are creating value."

The Polaris Program is based on quantitative research to identify pricing trends in global markets. Polaris uses a multi-model, multi-timeframe approach to invest in global commodity, interest rate, equity index and foreign exchange markets. The program looks at 30 years worth of data to identify pricing anomalies that are repeated over different time horizons. A typical holding period can last from 5 days to one year. The investment team at Rotella is constantly refining the dataset and technical models that support Polaris to ensure trading signal accuracy.

A new regime

The double-edged sword of constant research is getting so good at predicting success that you also know immediately when you might fail. Shortly after the financial crisis, Rotella's investment team concluded that equities were likely to outperform. Historically, managed futures have served as a strong diversifier for a traditional stocks and bonds portfolio during normal market regimes. But, Rotella's research predicted that an economic recovery spurred on by quantitative easing would make it difficult for managed futures funds to do well.

Rotella took the opportunity to retrench. The firm told outside investors in Polaris that while the trading program was going to continue, Rotella planned to focus most of its effort on research in order to understand the new normal. "We wanted to focus a lot of our research on the equities space," Prakasam explained. "We started to look more at position sizing and how market patterns were likely to shift."

From 2008 onward, Rotella's core focus was quantitative research. The firm expanded its office in Washington State, near Microsoft and Amazon, taking advantage of the local high-tech workforce. Rotella also seeded other hedge funds and CTAs during this time, trying to get a sense of what worked and what didn't. Most of it didn't. "What we found was that the value they were creating wasn't more than what we were already doing in-house, so we eventually exited those investments," says Ram.

Rotella's research process returned some interesting insights. The investment team learned how the new market regime could impact position sizing. What's more, analysts started to find new ways of leveraging artificial intelligence to extract value.

"We had looked at using neural networks to figure out pricing in the early 2000s, but we didn't think using neural nets for that purpose at that time was very profitable," Prakasam says. "What tends to happen with neural nets is that they aren't as transparent as a technical model. So if things are going well everyone is going to be happy with it, but if there's a drawdown we may not be able to explain it and that's a problem."

In order to deal with the 'black box' nature of some aspects of machine learning, Rotella has started defining narrow and specific use cases for AI within their research process. "AI, unfortunately, has become a buzzword," says Prakasam. "We don't think it makes sense to use AI for everything. Instead, we look at the problem we are trying to solve and determine the best use of machine learning to create value and help us with the problem at hand."

According to Prakasam, one of the ways they've been able to leverage machine learning is to help the investment committee handle unstructured investment data. As the amount of information about a given company grows through online content, social media and traditional means like earnings disclosures attempting to control the firehose is impossible for human analysts. Instead, the research team at Rotella have created an almost symbiotic relationship with their machines. For any company in the S&P 500, Rotella's analytics platform is learning to make bull and bear calls from a combination of new and old information. Whereas before analysts may have read a piece of news or an earnings transcript on a company and tagged it as bullish or bearish manually, the goal is to train the machines to be able to do that for each piece of data for each company in the index. Previously hand tagged information provides the training set for the machines. From there, the machine can analyze new data and the investment team oversees the process - checking the calls the machines come back with for accuracy.

"We don't have enough analysts in the world to analyze all of the stocks in the world and come up with a sentiment, but you can do this with AI. The ability to understand a lot of information is the edge," Prakasam contends

New Beginnings

Since 2008, Rotella has been busy refining its edge and adding to Polaris' track record. In 2017 the firm made Polaris available in a UCITS structure. "This is a big step for us. We are proud to be able to show that we have been assessed rigorously by regulators and we have also secured seed funding for that effort," Ram says.

The research also yielded a new investment strategy - Eta Carinae. The investment team launched Eta Carinae with internal capital in December of 2008. The strategy differs from Polaris in that it is more focused on equities, but retains Rotella's core expertise of identifying pricing movements across a global universe of securities. The models are a blend of ideas looking to capture trend, breakout and mean reversion behavior in the short and medium term time horizon.

"We just launched Eta Carinae as a fund last year and the reaction has been positive so far," says Ram. "We can show the track record from 2008 to now and we think it shows that we stand by our ideas to trade our programs with our own money through the family office first."

Eta Carinae is traded 24 hours a day, taking advantage of opportunities across time zones.

B. McCann

 
Today's Exclusives Today's Other Voices More Exclusives
Previous Opalesque Exclusives                                  
More Other Voices
Previous Other Voices                                               
Access Alternative Market Briefing


  • Top Forwarded
  • Top Tracked
  • Top Searched
  1. Gundlach's Forecast for 2019[more]

    Jeffrey Gundlach said that 2019 will mark the start of a period when bond markets must recon with the rising federal deficit. In his most passionate comments ever on this topic, he said the exploding national debt and liabilities involving pension funds, state and local government governments and So

  2. Institutional Investors: Texas Teachers allocates $1.9bn to 12 strategies, $53.6bn LGPS asset pool readies double global equity launch, The rise of co-investment SPVs for institutional investors, Wisconsin commits $500m to alternatives[more]

    Texas Teachers allocates $1.9bn to 12 strategies From PIonline.com: Texas Teachers Retirement System, Austin, committed a total of $1.871 billion to 12 private market investment strategies managed by existing managers, a transaction report showed. The biggest move by investment officers o

  3. News Briefs: The Jeff Bezos divorce: $136bn and Amazon in the middle, Bridgewater Associates partners with consulting firm, Cyprus no longer Mediterranean haven for Russian businesses, Goldman Sachs on course to launch cash management in mid-2020[more]

    The Jeff Bezos divorce: $136bn and Amazon in the middle From Mint: The announcement by Amazon founder Jeff Bezos, the world's wealthiest man, and his wife that they will divorce has captivated the imagination -- how will they split his giant fortune, estimated at $136 billion? And what wi

  4. Institutional Investors: Institutional investors plot large property allocations, Qatar Investment Authority aims to reach $45bn in U.S. investments, Louisiana Teachers assigns $200m, SoftBank move to slash WeWork investment sends shockwaves[more]

    Institutional investors plot large property allocations From FT: Institutional investors plan to make large allocations to property in 2019, taking them closer to their desired targets. At least €72.4bn of new capital is expected to flow into real estate this year, according to a

  5. Legal: Attorney sues after Tampa hedge fund goes under, Hedge funds showing increased interest in litigation claims, Argentina sued again by hedge fund on bonds tied to GDP growth[more]

    Attorney sues after Tampa hedge fund goes under A Tampa hedge fund company may be in legal trouble after the firm that was holding its money was sued by a lawyer who claimed "gross negligence" caused the fund to lose tens of millions of dollars. James Cordier, the head of optionseller