On some strange features in Kepler EB light curves.

The Kepler Eclipsing Binary Catalog Third Revision contains data for about 3000 systems (Prša et al, 2011). Enumerate all the papers connected to the database we could get around 1300 papers in the literature published during the last decade, so its impact is definitely huge. One could suppose that this uniquely large database with unprecedented accuracy would provide a good foundation for detailed studies as it was advocated, e.g. in Wilson (2003), to look for an efficient, effective, and automatic processing of massive database secured by large-scale surveys but ‘Eclipsing binary (EB) modeling naturally partitions into several areas, headed by the overall issue of how to find astrophysically useful numbers.’ (Wilson, 2006) However, the number of studies discussing astrophysical details is much more limited. Pursuing this main goal, we have browsed through the catalog of corrected and detrended light curves in the database, and have found a few typical features created during data processing. These detections demonstrate that one must be careful around the details before using the data directly to deduce any physical conclusion. The artificial side effects contaminate the real data and can mimic otherwise pretty intriguing real effects.