Bob Pardo taught himself computer programming as he anticipated the direction of trading back in the 1980s when he was in charge of floor operations for Salomon brothers and worked on supporting the execution of trading legend John Meriwether’s hedge fund.
“I bought an Apple II [and] taught myself programming. I wanted to use it to create trading signals and do market analysis,” Pardo says.
This led him to build trading software, and when he was able to sell $3,000 worth of software at a trading show — a lot of money in the early 1980s — he saw there was interest and launched a software business.
“I built a program that became one of the first really capable and affordable charting programs for the small investor: Chartist. It evolved into Advanced Chartist, which then evolved into Advanced Trader,” Pardo says. “I built a program into one of the first charting programs for the small investor — The Chartist. Later it evolved to Advanced Chartist, which then evolved into Advanced Trader.
At the time it became popular to sell off-the-shelf trading signals (see “Trading System Shopping,” MT June 2018), but Pardo had something altogether bigger in mind.
“With Advanced Trader we built a programming language that allowed people to do testing of trading ideas. This was before Easy Language,” Pardo says. “We evolved that into Expert Trader that allowed people to test a portfolio of strategies in very flexible ways. I was always pushing the envelope for learning new stuff.”
His work gained the attention of a big name. “Someone at Goldman Sachs found us and they were looking for an in-house white label program that could be adapted to their needs,” Pardo says.
Unfortunately, when the three-year project ended, their benefactor at Goldman had passed away and his replacement was not interested in continuing the project, but Pardo’s team had made progress. “We expanded Advanced Trader to be real-time and adapted it to read their data feeds, and we built an API into it,” he says.
Pardo would then work with Daiwa Securities. Traders there were able to produce their own trading signals through the family of models Pardo produced.
One of Pardo’s models stood out and he continued to develop it into what would become his XT99 model that he perfected by the early 1990s, and would be the basis for a CTA. After meeting Pierre Tullier of Dunn Capital Management, Pardo agreed to work with him in offering the strategy to Dunn investors. Dunn ran it through its proprietary risk measures and would raise money for the program.
The strategy was highly successful, earning an annual rate of return of 21.07% from June 1999 through August 2011. In 2008 alone, XT99 earned 142.01%.
Pardo was happy earning fees from the strategy that managed roughly $40 million at its peak, but Pardo says that Dunn did not aggressively market the program that was highly correlated with Dunn’s in-house strategies.
XT99 was an intermediate to long-term breakout strategy that fit into the broader CTA trend following space. “It traded off of the primary cycle of all its markets,” Pardo says.
A unique aspect of XT99 is that while it was the same strategy for all markets in terms of the general parameters, he customized it to each of the 40 markets it traded.
After the financial crisis, and with the world of managed futures becoming more institutional, returns like the onesXT99 produced — four years of greater the 60% including 142.01% in 2008 — became more of a liability than a selling point. “People literally complained about our 2008 performance; it was actually a turn off,” Pardo says.
He would work on lowering the volatility and stopped offering it after August 2011, despite being in the midst of a positive year.