Doug, the devil is in the details. Models are very helpful when the programmer who writes the modeling software has an accurate description of the process being modeled by the computer. For airplanes and nuclear weapons, we understand the process really well and understand the margins of error acceptable for modeling them. With climate change, the process of climate is not well understood and some factors are left out of the models altogether. When that happens, the programmer does not have a sufficiently accurate understanding of what is being modeled and the resulting software becomes much more suspect. If the program is off even a little bit in tolerances, the resulting predictions can be way off in the outcomes the models project.