Kepler Planet Detection Metrics: Robovetter Completeness and Effectiveness for Data Release 25

In general, the Kepler pipeline identifies a list of Threshold Crossing Events (TCEs), which are periodic flux decrements meeting certain criteria (Jenkins, 2017). These TCEs are reviewed and those that appear consistent with astrophysically transiting or eclipsing systems are classified as Kepler Objects of Interest (KOIs). Further review is given to KOIs, which are then dispositioned as Planet Candidates (PCs) or False Positive (FPs). FPs are further denoted by four major flags that indicate if the signal is Not Transit-Like (NTL), due to a Stellar Eclipse (SS; previously referred to as Significant Secondary), and/or due to contamination from a source other than the target as evidenced by a Centroid Offset (CO) oran Ephemeris Match (EM) with another object. This entire TCE review process is known as dispositioning or vetting.In the first five Kepler mission planet candidate catalogs (Borucki et al., 2011a,b; Batalha et al., 2013; Burke et al., 2014; Rowe et al., 2015), TCEs were manually examined on an individual basis and dispositioned using various plots and quantitative diagnostic tests (see e.g., Coughlin, 2017). In the sixth catalog, Mullally et al. (2015a) employed partial automation via simple parameter cuts to automatically disposition a large fraction of TCEs as not transit-like. Mullally et al. (2015a) also used an automated technique known as the centroid Robovetter (Mullally, 2017) to automatically identify some FP KOIs due to centroid offsets - a telltale signature of light contamination from another target. The remaining targets were manually dispositioned. In the seventh catalog, Coughlin et al. (2016) automated theentire dispositioning process using what is collectively known simply as the Robovetter.In the eighth and final mission catalog, Thompson et al. (2017) use a revised Robovetter to automate the dispositioning of all TCEs with an emphasis on creating a catalog suitable for accurately determining planet occurrence rates. In order to calculate accurate occurrence rates, the completeness and effectiveness of the Robovetter must be characterized. We define these terms as applied to the Robovetter, following Thompson et al. (2017), as:1. Completeness: The fraction of transiting planets detected by the pipeline that are classified as planet candidates by the Robovetter.2. Effectiveness: The fraction of false positives detected by the pipeline that are classified as false positives by the Robovetter.The remainder of this document describes products that can be used to quantitatively assess Robovetter completeness and effectiveness for an arbitrary set of Kepler stars.