In fact, the researchers found that the expert handicappers used models that were far more complex than what the nonexperts used, so-called multiplicative models in which the values of some factors (such as track condition) altered the importance of others (such as last-race speed).
