I just saw Bence's writeup about the Excess Uncertainty Explorer and I added the plugin (note, I had to change the reference material in the code from Z_91500 to Z_GJ1 before it would properly open).
I noticed something strange. For zircon U-Pb, we find that the AutoSmooth spline always looks like a good fit to the standard reference data, but when I open the Excess Uncertainty Explorer tool, the spline doesn't fit the same way at all. When I view the DC206_238 data in Time Series, I see nice gradual curves as it fits the data, but in Excess Uncertainty Explorer the DC206_238 spline shows up as just a straight line much like the LinearFit spline, and thus yielding much larger jack knifed values. I don't seem to be able to add screenshots to this post to demonstrate this.
Is this likely an artifact of the Uncertainty Explorer tool, or could it be that using the AutoSmooth spline is causing the propagated errors to be calculated wrong?
Thanks for any comments,
Lisa