RCtest 1.1.0

• estimate_forecast_variance() now requires a realized argument. Forecast errors are computed relative to the realized outcome instead of the benchmark forecast.
• Fixed a redundant variance division in superior_predictive_ability_test(), white_reality_check_conditional(), white_reality_check_cdf_approx(), kullback_leibler_test(), and reality_check_zp_test(), which inflated test statistics by a factor of roughly sqrt(sample size).
• Fixed white_reality_check_cdf_approx()’s bootstrap null-cenetring, which previously caused the test to reject H0 in almost all samples regardless of the data. This is a separate fix from the one above.
• Fixed the one-sided tail direction in compute_per_model_statistics() (H1 = "more"/"less"). Two-sided results are unaffected.
• Minor editorial corrections to package description text.