Since that time, Mr. Decker has continuously developed and refined his tools. His models use the best propagation algorithms available plus a variety of unique code modules, including the stepped radial method [G.A. Hufford & S.J. Paulson] CSPM models...for most efficient use of area in computer predictions, rounded obstacle diffraction [Gerks & Vogler] with rapid term convergence, optimal containment of clutter [M.A..Weissberger, C. Lava], including JTIDS study results, tree heights, and density, option to integrate ray path through clutter using the method of Lava.

 The propagation algorithms used have progressed from Longley-Rice (and upgrades), through a unique "Indirect Method" of selecting and using diffracting objects from an ensemble of objects along a path, to what they are today using the CRC-PREDICT^TM [J.H. Whitteker] quasi-exact two-dimensional Fresnel-Kirchhoff wave-front method employing a chained calculation (in range) of superimposed specular reflection and diffraction computations based on Huygen's Principle.

 The current model, Athena32,  is Whitteker fine-tuned based on recent R-data analyses, and incorporating the LULC algorithms of K. Low.

 Terry Cory of Wave Concepts has known Mr. Decker since the inception of his work, and has collaborated with Mr. Decker in several land-mobile system studies for Spectra Associates. Since forming an association with Mr. Decker in 1996, Wave Concepts personnel have been instrumental in developing what is now the premier propagation and simulation tool in the world.

 CRC-PREDICT^TM (Witteker) conservatively claims a standard deviation of prediction error for the Fresnel-Kirchhoff code to be 7-8 dB using 500-meter resolution terrain data that characterizes the CRC-PREDICT^TM code implementation in the Canadian digital television broadcast service. Whitteker stated in his 1994 Austin paper that prediction sigma of 5 dB was noted in a smaller measurement sample. Decker has performed an exhaustive analysis of credible path measurements performed by ITS over a period of years in the western U.S., and in Virginia. The results of these analyses show sigma's averaging 5-6 dB based on 3-arc-second terrain data and 250-meter LULC data; some paths being much better. These validations have convinced Decker that in environments having "mild" clutter in terms of morphology, the true prediction sigma is at least as small as 5 dB and in the limit may well be as small as 3.6 dB neglecting measurement error. The fact is that this technique is the only one which accurately predicts location variability due to diffraction in the near field regions of the base station and diffracting obstacles according to a theory satisfying the boundary conditions imposed by a chained calculation in range. Regardless of the terrain data base resolution, the model is the only one that can resolve variations to a resolution of less than about 200 meters (such as, 100 meters or 30 meters or 10 meters) providing the clutter profile has been adequately described within a similar resolution.