We present a theoretical study of the complex relation between radon and its long-lived progeny implanted in glass surfaces. The well known (extended) Jacobi room model, which is normally used to describe radon and its progeny in a room, was transformed into a two-parameter model revealing a linear correlation between long term radon exposure and surface activity due to implanted radon decay products. Furthermore, this new approach made integration into a Monte Carlo simulation possible so that the large variation of different room model parameters could be taken into account. This allowed the calculation of a probability distribution for radon exposure from the measurement of the implanted 210Po activity. The availability of a 95% confidence interval for the radon exposure is valuable in the application of retrospective radon assessment in epidemiological studies.