Sоme lаrge cоmpаnies, including Gоogle, hаve experimented with internal prediction markets, where employees trade on forecasts about real business outcomes. This mirrors Best Buy's real TagTrade market from this week's assigned reading, where the crowd's aggregated forecast averaged 99.5% accuracy versus paid experts' 5% error rate on the same forecast. What is one real limitation of the wisdom of crowds effect, based on the research?