https://foursevenfive.com/blog/gutex-vs-xps-insulating-in-a-hot-house-world/

So this is really interesting. This company constructed two identical boxes, one with Glutex wood fiber insulation rated at R-5.7 and one with XPS foam insulation rated at R-10 and monitored the internal temperatures over a month. And if you are still reading you’ll probably guess what happened, the Glutex wood fiber insulation performed better, even though it is almost half the r-value of the XPS foam!

Part of the issue is how the r-value is determined:

Standard tests for R-value use a steady-state temperature (ASTM 2018), but our enclosures exist in a world of highly variable temperatures. Furthermore, R-value tests typically only include conductive heat transfer, missing any influence of radiant heat transfer, and are conducted under unrealistic conditions, such as 50 degrees F at the interior with 100 degrees F at the exterior (Bailes 2013).

This scenario actually happens almost all the time. Someone notices a difference in how things perform (in this case insulation) and comes up with a way to measure the difference (the r-value). This method of measuring the difference is almost always optimized to be cost effective to test which means it usually is quick to do and/or is based upon a limited number of variables. (i.e. testing at a steady state temperature, ignoring radiant heat) Based upon this new way of measuring the difference, companies come up with new products that are optimized for profit based upon that measurement (i.e. highest r-value/inch) and given enough time, these products take over the marketplace, even though they do very poorly outside of the test conditions.

Another great example of this is Polyisocyanurate insulation which has the amazing feature of being less effective at resisting temperature the colder it gets!