A critical value is the threshold with which a test statistic is compared in hypothesis testing. If the statistic falls beyond the critical value, in the region that would be unlikely were the null hypothesis true, the null is rejected. The value depends on the chosen significance level, on whether the test is one-sided or two-sided, and on the distribution the statistic follows.
For a statistic that follows the standard normal distribution, the two-sided critical value at the 5 percent level is 1.96, and the one-sided value is 1.645. For small samples the values come from Student's t-distribution and are larger. A strategy whose mean return has a t-statistic of 2.4 exceeds 1.96, so the hypothesis of zero mean return is rejected at 5 percent. Some tests, such as unit root tests, have their own tables. When many strategies are tested, researchers raise the critical value to limit false discoveries.