From 765b31724508ad4c243cd7f3196d0edf33bfbdae Mon Sep 17 00:00:00 2001 From: arvindv92 Date: Tue, 10 Nov 2020 10:59:04 +0100 Subject: [PATCH] Update statisticaltests.rst Fix what looks like a typo in line 301. For the dataset N_B, the likelihood ratio should be small (not large), since the numerator is small and the denominator is large. Also, thanks for this workbook. It is very helpful --- docs/source/statisticaltests.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/source/statisticaltests.rst b/docs/source/statisticaltests.rst index dcf235c..e1eefa5 100644 --- a/docs/source/statisticaltests.rst +++ b/docs/source/statisticaltests.rst @@ -298,7 +298,7 @@ and :math:`\vec{N_S}|H_{B})` will be small, since the data is improbable under t hence the ratio will be large. Conversely for a dataset :math:`N_B` that is very background-like :math:`L(\vec{N_B}|H_{S+B})` will be small, since the data is probable under this hypothesis, and :math:`L({\vec{N_B}}|H_{B})` will be large, since the data is improbable under this hypothesis, -hence the ratio will be large. +hence the ratio will be small. With a likelihood-ratio test statistic, frequentist :math:`p`-values can be calculated for observable data distributions or arbitrary complexity as the test