A method and system of reliably detecting a reactive jamming attack and estimating the jammer's listening interval for exploitation by a communication system comprises channelizing one or more signals of interest (SOI), channelizing one or more signals of unknown origin (SUO), identifying frequency support patterns for the SOI and SUO using Bayes thresholds, comparing SOI and SUO detection map histories, and determining a percent match, where a match percentage above a specified minimum indicates a reactive attack. Edge detection can be used to enhance jammer support. Embodiments further detect reactive jammer adaptation to changes in the SOI's frequency support. Embodiments include detectors that are insensitive to jammer modulation and/or signal type. A jammer reaction delay and/or size and periodicity of receive window can be detected. Embodiments determine if a jammer is copying and retransmitting the SOI's waveform(s), and/or if the jammer is anticipatory.