I am having trouble using CSAPS function. Consider a slightly modified example from matlab manual
x = linspace(0,2*pi,21); y = sin(x)+(rand(1,21)-.5)*.1;
pp = csaps(x,y, .4)
Changing the smoothing parameter p, which is equal 0.4 in the case above, gives you different fitting curves. However, if I extend the range of data sites, say x=linspace(0,100*pi,21), then no matter which value of smoothing parameter p I use it gives me exactly the same curve. Conversely, if I shrink the interval then it gives me just a straight line irregardless of the p-value. Any thoughts?