Ambiguity dilution of precision (ADOP) and position dilution of precision (PDOP) are two popular scalar-diagnostics used in Global Navigation Satellite System (GNSS) positioning. Where the ADOP is a predictor for carrier-phase ambiguity resolution performance, the PDOP is meant to predict the receiver-satellite geometry's capability for precise positioning. We will show, however, that although the PDOP works well for code-based positioning, one has to exercise great care in using the PDOP for real-time kinematic (RTK) positioning. We show that the ADOP and PDOP have distinct behaviors, an important consequence of which is that one can have time periods with small PDOPs, and thus seemingly good geometry for precise positioning, but at the same time large ADOPs, thus showing that successful ambiguity resolution and therefore precise positioning will not be possible. Also, the reverse situation may occur, i.e., having large PDOPs with small ADOPs. In such a situation, the large PDOPs should not automatically lead to the conclusion of poor position performance, because the large gain that ambiguity resolution brings will often still make precise positioning possible. We will analyze and explain this complementary behavior of the PDOP and ADOP and demonstrate this both analytically and empirically. For this analysis we use real Global Positioning System (GPS) single- and multifrequency signals and GPS/Quasi-Zenith Satellite System (QZSS), GPS/Navigation with Indian Constellation (NAVIC) L5 signals of two baselines located in Perth, Australia.@en