Well, I blew it yesterday, reporting a "bug that wasn't a bug", just careless reading on my part. But this one seems "more real" to me ...
In the Solving Tutorial, Task 2-3 shows Nonlinear Least Squares Fitting, a nice example fitting 12 X-Y data points to a Weibull distribution. Very nice.
However, Step 7 says "Calculate the mean squared error", calculated by dividing the sum-of-squares of the residuals by n-2 (number of points minus the number of parameters). The problem is that n is calculated as "length(y)-1", whereas I think it should (logically and mathematically and MathCad-mathically) be simply "length(y)" (which, as I verified, evaluates to 12).