The double dose resonse curves are very difficult to fit and many times a nonlinear least squares type of algorithm will fail. This application take a three stage approach to find the best possible fit starting with a guess based on the data ranges, followed by a brute force search, and finally a nonliner squares fit where possible.