Improvements to existing transit detection algorithms and their comparison

In Tingley (2003), all available transit detection algorithms were compared in a simple, rigorous test. However, the implementation of the Box-fitting Least Squares (BLS) approach of Kovacs et al. (2002) used in that paper was not ideal for those purposes. This letter revisits the comparison, using a version of the BLS better suited to the task at hand and made more efficient via the knowledge gained from the previous work. Multiple variations of the BLS and the matched filter are tested. Some of the modifications improve performance to such an extent that the conclusions of the original paper must be revised.