Prediction skills of real-world market timers

The conventional approach for describing a market timer's performance is to compute the annualized return and volatility of his portfolio. These descriptors, however, do not reveal how much of the timer's performance is due to his prediction skills and how much is due to luck. In this article, the performance records for a group of twenty-five real-world timers are transformed into different set of performance descriptors that actually do define the timer's ability to accurately predict bull and bear markets. These prediction skills also provide a framework for comparing the skill of the real-world timers with a coin-flipping timer as well as with a buy-and-hold investment strategy. This comparison shows that the timers are, on average, somewhat better than a coin-nipping timer but are not likely to beat a buy-and-hold strategy, particularly after taxes and management fees are considered.