Peter Reinhard Hansen proposes a new test for superior predictive ability (SPA), which is more powerful and less sensitive to poor and irrelevant alternatives compared to the reality check (RC) method. The new test uses a studentized test statistic and a sample-dependent null distribution to improve its performance. The theoretical analysis reveals that the RC can be manipulated by including poor and irrelevant forecasts, while the new test addresses this issue by standardizing the test statistic and using a sample-dependent null distribution. The advantages of the new test are confirmed through Monte Carlo experiments and an empirical study comparing regression-based forecasts of U.S. inflation to a simple random-walk forecast. The results show that the random-walk forecast is outperformed by regression-based forecasts, with the best sample performance achieved by models with a Phillips curve structure. The article also discusses the bootstrap implementation of the test and provides a detailed explanation of its finite-sample properties.Peter Reinhard Hansen proposes a new test for superior predictive ability (SPA), which is more powerful and less sensitive to poor and irrelevant alternatives compared to the reality check (RC) method. The new test uses a studentized test statistic and a sample-dependent null distribution to improve its performance. The theoretical analysis reveals that the RC can be manipulated by including poor and irrelevant forecasts, while the new test addresses this issue by standardizing the test statistic and using a sample-dependent null distribution. The advantages of the new test are confirmed through Monte Carlo experiments and an empirical study comparing regression-based forecasts of U.S. inflation to a simple random-walk forecast. The results show that the random-walk forecast is outperformed by regression-based forecasts, with the best sample performance achieved by models with a Phillips curve structure. The article also discusses the bootstrap implementation of the test and provides a detailed explanation of its finite-sample properties.