Radio channels are usually modeled as a well‐defined Rice process. The statistics of its wrapped phase (i.e., phase values in [−π,π)), such as the mean, variance, and probability density function (pdf), are known. The absolute phase, i.e., the accumulated phase change over an observation interval, is considered here as a new channel variable. Its use for channel characterization can extend to cognitively track users. However, there is very little knowledge about the statistics of the absolute phase. In fact, the known, associated effects of the absolute phase, such as various click noise contributions, are not consistently treated or interpreted in the literature. The definitions of absolute phase, based on both unwrapping and on other methods previously discussed for FM receivers, lay a basis for analysis of the mean, variance, and pdf of the absolute phase for a well‐defined Rice process. The conditions are identified for approximate pdf models to hold, and it is noted that pdfs for the absolute phase for small or medium Rice factor and small observation interval are open problems. Simulations are used to support the analysis and discussion. Copyright © 2009 John Wiley & Sons, Ltd.