Bias and consistency in time delay estimation methods: case of the double quasar HE 1104-1805

Abstract

We present a short re-evaluation of a recently published time delay estimate for the gravitational lens system HE 1104-1805 with emphasis on important methodological aspects: bias of the statistics, inconsistency of the methods and use of the purposeful selection of data points(or so-called "cleaning") at the preprocessing stage. We show how the inadequate use of simple analysis methods can lead to too strong conclusions. Our analysis shows that there are indications for the time delay in HE 1104-1805 to be between -0.9 and -0.7 years, but still with a large uncertainty.