A simple method of sample size calculation for unequal-sample-size designs that use the logrank or t-test.