This article, by Steven Few, does a great job of explaining when to use colors when making graphs, and how to choose which colors to use.

Prism handles perfect data sensibly when you choose an ordinary nonlinear regression fit (not robust, no request to count outliers). Prism reports 'perfect fit' and states accurate values of the parameters. It also reports that the sum-of-squares is 0.0 and R² is 1.000. It does not report values for the standard errors of the parameters or for their confidence intervals. 

If you choose a robust fit (or choose to count or exclude outliers, which requires a robust fit first), Prism gets confused with perfect data.  It reports that the fit was "interrupted" and reports "value too large" instead of parameter values. If your data are perfect, it doesn't help to do a robust fit and doesn't make sense to look for outliers. Turn off these options, and Prism will fit the curve just fine. We'll fix this glitch in 5.03 and 5.0c. 

Much of the battle when learning statistics is understanding terminology. The terms logistic is confusing because it has three meanings which have little relationship to each other: logistic population growth, logistic dose-response curves, and logistic regression.

Read more.