U.S. markets close in 1 hour 50 minutes
  • S&P 500

    4,135.80
    +7.81 (+0.19%)
     
  • Dow 30

    33,628.33
    -117.07 (-0.35%)
     
  • Nasdaq

    13,949.81
    +99.81 (+0.72%)
     
  • Russell 2000

    2,216.28
    -17.50 (-0.78%)
     
  • Crude Oil

    60.06
    +0.36 (+0.60%)
     
  • Gold

    1,745.40
    +12.70 (+0.73%)
     
  • Silver

    25.39
    +0.52 (+2.08%)
     
  • EUR/USD

    1.1946
    +0.0028 (+0.24%)
     
  • 10-Yr Bond

    1.6290
    -0.0460 (-2.75%)
     
  • GBP/USD

    1.3747
    +0.0005 (+0.03%)
     
  • USD/JPY

    109.1000
    -0.2760 (-0.25%)
     
  • BTC-USD

    63,516.10
    +3,539.80 (+5.90%)
     
  • CMC Crypto 200

    1,359.74
    +65.75 (+5.08%)
     
  • FTSE 100

    6,890.49
    +1.37 (+0.02%)
     
  • Nikkei 225

    29,751.61
    +212.88 (+0.72%)
     

The ‘stonks’ market caught the A.I. algorithms off guard, too

Jeremy Kahn
·3 min read

It’s widely known that last month’s surge in beaten-down stocks such as GameStop and AMC that were championed on Reddit’s WallStreetBets community caught a number of well-known short-sellers flat-footed.

But it seems savvy traders weren’t the only ones surprised by the market’s wild moves. It turns out the machine brains didn’t see it coming either.

Hedge funds that depend on artificial intelligence to make their trading decisions were also seemingly unable to figure out how to play the stonk market, contributing to a dismal start of the year for these investors, according to figures from Eurekahedge, a company that tracks the performance of different hedge fund strategies.

The Eurekahedge AI Hedge Fund Index declined 3.38% in January, its worst one-month performance since the index’s inception in 2010. It’s also worse than the January 2021 performance of any other hedge fund strategy Eurekahedge tracks and worse than the overall performance of the S&P 500, which slid 1.11%. The Nasdaq, meanwhile, gained 2.9% for the month.

Mohammad Hassan, Eurekahedge’s chief analyst, said that while most of the funds that make up the AI Hedge Fund Index were flat or slightly down for the month, the index’s poor performance was driven by a few funds that were hammered, experiencing double-digit percentage declines, possibly because of exposure to U.S. small-cap stocks.

Many hedge funds use A.I. algorithms that are trained on historical market data. These systems can perform very well in periods when trading conditions are relatively stable. A study from research firm Cerulli, published in May 2020, found that A.I.-driven hedge funds produced cumulative returns of 34% over the previous three years, compared to just 12% for hedge funds overall during the same period.

But when there is a sudden change in market conditions, such as an unexpected interest rate decision or political event—or, yes, a Reddit-crazed army of retail investors YOLO-ing the price of a few select equities to the moon—those A.I. algorithms may not fare so well.

While lots of trading firms use machine-learning algorithms of various kinds to sift market data for possible profitable correlations, and plenty more use A.I. algorithms to execute the trades that their human analysts dream up, what sets apart the 16 funds currently in Eurekahedge’s AI Hedge Fund Index is that they have largely handed trading authority over to the algorithms. The machines are calling the shots, for good or ill.

“We have documented a series of events like this over the years, such as the Brexit referendum vote and the Greek referendum and Trump’s election, where something unusual is happening that is driving the market, where these strategies get caught off guard,” Hassan says.

Yes, sometimes the smart money gets beaten by a gang of self-described ’tards. And sometimes artificial intelligence is plenty dumb.

More must-read finance coverage from Fortune:

This story was originally featured on Fortune.com