Anomalous frequency spikes derived from complex-trace analysis have long been considered more as a flaw of algorithms than a valuable seismic attribute. However, synthetic models and field data show that instantaneous frequency spikes are clearly linked to physical models that could help geologic interpretations of seismic data. Frequency spikes are related to relative amplitude minima along a seismic trace that occur either at the merging point of the top and base reflections of a wedge (type I) or at the tip of the wedge (type II). A type II spike is indicative of a seismically thin bed (average or thinner in this study; λ denotes wavelength), although the bed thickness tends to decrease with the data frequency and bed spacing in a multibed acoustic-impedance profile (in the range of to in this study). Interpretive values of instantaneous frequency include recognizing thin reservoirs that are otherwise difficult to detect, mapping relative thickness changes of lithofacies, detecting lithofacies boundaries and faults, and locating a stratal discontinuity or flow barrier. Uncertainties caused by type I spikes, noise-derived fake spikes, and the lack of lithologic and time-stratigraphic information can be reduced by integrating well data, local geologic models, lithology-indicative seismic attributes (e.g., 90°-phase data), and facies-oriented seismic displays (e.g., stratal slices).