I am using the routine mrqmin for modelling the data. My initial data is that of radial distribution function & the model function which is being used is A*exp(-B*r(*sin*C*r). Here A,B,.C are three unknown parameters. I don't know hat are the appropriate guesses for them . I am just feeding them randomly. The value of chi-square is reducing. Bur it's never less than 100 for this function. The guesses are returning in improved form. But when i put them in the above function & plot it, function doesn't give the exact shape of radial distribution function. It doesn't fluctuate around 1. I have also tried for the model fuction A*exp(-B*r(*sin*C*r)+1. Now, the fn. fluctuates around 1. But, the results are not satisfactory. How can i overcome this problem. Also how will i come to know what are the expected outcomes of the guesses.