In this paper the ability of a variety of backtesting experiments to identify a model with misspecified volatility is examined. This quantitative testing assumes five years of risk factor observations, considers overlapping and non-overlapping backtest observations with horizons out to a year,...