Collins started out writing code and as the business grew, he covered more traders and learned the process of exchange for physicals (EFPs). He would bring this knowledge to Dunn, which was a customer of Geldermann and had grown quite large.
“Bill and Pierre [were not using] EFPs. I came down here in the early ’90s and said, ‘Hey guys, you are getting killed in the way you are putting your orders in. You are getting too big and people are taking advantage,’” Collins says.
In addition to working with Dunn to help manage its growth throughout the 1990s through his position with Geldermann, Collins would develop trading models and have Pierre break them down. “Finally, around 1999 [Pierre] said, ‘You know how to tear apart models. You know what is important, and you know where the weaknesses and strengths are in systems,’” Collins says.
In 2004, Collins was working for a family office when Pierre decided to retire and suggested they work on building a trading system together. “Pierre didn’t want to compete against Dunn. We stayed away from futures (and trend-following). Finally, in 2008 I wrote the first program that is part of the current Ocean program,” Collins says. “[Pierre] was flabbergasted at how robust it was. He was adamant that we start trading as soon as possible.”
Unfortunately, in early 2009 Pierre was diagnosed with stage 4 liver cancer. “That took us all by shock. Basically, he was handed a death sentence,” Collins says.
Tullier died before the short-term program launched but played a major part in its creation. Collins launched the CEC Capital Ocean Fund in late 2010 and it has produced a compound annual return of 21.61% in just under two years, with a Sharpe ratio above 1. The program is managing $7.1 million and is up 10.5% year-to-date through June.
“It is a multi-strategy approach to trading the S&P 500 on a short-term basis. It [includes] pattern recognition, mean reversion, volatility breakout, momentum, trend-following and countertrend-following,” Collins says. The program launched with three models and he added pattern recognition and countertrend models in May 2011 after a drawdown; he more recently added a trend-following model.
“We are diversified by strategy. The CTA industry is much more advanced than the equity industry in terms of empirical measurements and facts about correlation. Who needs mid-caps, small-caps [and] value when they all tank when the markets tank. We need diversification by strategy because as we have seen through experience, non-correlated assets when you really need them to be non-correlated tend to correlate,” Collins says.
The average holding period for all models is one hour, and most trades are never carried overnight. He is confident in his models because they have been tested rigorously based on the expertise of the best minds in the industry. “I was very fortunate to have Pierre as a mentor,” Collins says. “I learned to find weaknesses. I know how to tear apart systems because I was taught how to tear apart models.”
Collins continues to utilize those skills with plans to launch a diversified trend-following program in August. He says the new program will be focused more opportunistically and will not look like traditional trend programs such as Dunn.