A Blackstone-Backed Quant Gains 20% by Chasing the Money Flows

(Bloomberg) -- A machine-learning hedge fund backed by Blackstone Group Inc. is enjoying a growth spurt after notching a 20% gain in this year’s wild pandemic markets.

Bayforest Capital Ltd., which employs just five people in London, is set to oversee $235 million in managed accounts over the coming month compared with $45 million at the end of 2019. It also plans to launch a fund for institutional investors later this year.

A record of positive gains every month in this year’s cross-asset roller coaster is drawing fresh client attention to the firm run by Theodoros Tsagaris, a quant with previous stints at Tudor Investment Corp., GSA Capital and BlueCrest Capital Management.

He credits Bayforest’s success to algorithms surfing fast shifts in capital flows in real time. With a system trading futures based on the behavior of different investors and an average holding period of just eight days, the portfolio has managed to make money even as markets swing from despair to exuberance.

“We’re receiving billions of data points every day and we adapt our algos based on the new information,” said Tsagaris in an interview.

He aims to oversee $750 million by the end of the year, an ambitious goal for a firm that started trading in March 2018. But that looks somewhat more attainable after the $6 billion Blackstone Alternative Multi-Strategy Fund last month announced it’s allocating to Bayforest, joining sub-advisers including D.E. Shaw and Two Sigma.

A Bayforest strategy of following capital flows holds particular appeal in a volatile market jumping from record losses to historic gains. That’s especially so for systematic investors whose directional bets have been crushed by rapid shifts in bull and bear regimes. Investing styles that follow flows can range from trading ahead of index rebalancing to buying shares touted on websites followed by day traders.

Fast Money

Bayforest is also among a short list of quant funds this year that attribute their triumph to shorter holding periods and more adaptive trading signals, even as skeptics counter that constantly churning portfolios may not pay off in the long run.

Tsagaris has academic training on harnessing computing power to decipher noisy markets, having written his doctorate dissertation at Imperial College London on building adaptive algos from data streams.

He says Bayforest’s advantage is its ability to continually suss out new patterns, allowing it to adapt even when market participants try to hide their tracks by changing how they trade.

“What we are seeing is amazing alpha in areas that were not evident before,” he said, referring to the shifts in capital flows in both investing apps like Robinhood and exchange-traded funds.

©2020 Bloomberg L.P.