Some of the biggest commodity hedge funds are shutting down despite a rally in raw materials during the past year, according to a recent article in The Wall Street Journal.

In 2017, the article reports, “closures of commodities hedge funds outnumbered launches for the first time in data going back to 2000 ” (data from Eurekahedge) and the trend has continued this year. Fund managers and traders say it’s happening because “investors who were burned by the severe two-year market rout that started in 2014 aren’t rushing back” even though commodity prices are rebounding to multiyear highs. Fund closures also reflect an evolving market, the article explains, “one that is increasingly driven more by algorithms than fundamental information.”

Computerized trading has been slow to infiltrate the commodities market, which has typically involved transactions by phone as well as physical delivery and storage. But the article reports that “automated trading in energy-related contracts for the first time accounted for more than half of futures volume from late 2014 to late 2016,” indicating a shift in the sector.

But quant models alone may not be the answer. Eric Armitage, who started building such models for BP PLC in 2001 and now heads London-based East Alpha, said, “You can’t do this as a dark room quant,” adding that traders have to “know where the holes are in this data or where this data could be fake.”

Jonathan Goldberg, founder of energy-focused hedge fund BBL Commodities, echoes the argument, saying that “trading exclusively on information that everyone has access to, such as government data and inventory reports, is a fool’s errand.”