I can well believe that the algorithms doing the rounds were flawed in their assumptions, as that is what brought LTCM so close to collapse.  They discounted the "flight to quality" phenomenon during the 1987 Asian currency crisis, so got caught heavily betting the wrong way.  

I find it reassuring that even Nobel Laureates and former deputy governors of the Federal Reserve and Big Swinging Dicks from Wall Street can get it terribly wrong sometimes.  If the world were entirely predictable to the right formula, it would take the fun out of markets.