Hi ehuesman,
Here’s one more forum article which has a more simplified explanation of the durability difference between Dunlop and blended and all natural Talalay. Interestingly enough I found this in a google search because I didn’t list it as a reference post when I first wrote it quite some time ago. It’s not as technical as some of the others but it’s perhaps it’s more clear because of its simplicity.
I certainly understand this as well and if it’s a “fault” then I share it and your tendency to go into finer and finer grained layers of detail out of interest alone even though in many cases it’s past the point of the law of diminishing returns (and in some cases the time I have available compared to research that will benefit more people) and the raw data to get more and more specific is sadly lacking. In many cases I’ve spent hundreds of hours in deep searches on some topics over the years in efforts to find missing information that can help connect more data points.
In a site like this which is more about helping the majority to find a better mattress than they otherwise would … the amount of technical detail can already be overwhelming and it’s important to me to find a balance between too much information and not enough … both of which can lead to less then ideal choices. If anything I have crossed the line into too much (for the majority of people) on many occasions which in some cases can have the side effect of encouraging an undue focus on technical specs without the complete context to fully understand them at the expense of personal experience in testing mattresses.
Sometimes too, articles like the technical blog also need to be questioned according to what they are being compared with. You can see for example on the Latexco International site here that some of the specs for their Dunlop latex don’t agree with the specs of the Dunlop latex that Latex International was using for its comparison and they are also using different test methods.
You can also see in this Latex International document which has more detail yet than their technical blog that g1981c linked earlier that some of the information or underlying assumptions are also questionable.
For example they compare talalay cell wall structure and thickness at the same density as Dunlop instead of roughly the same ILD. While this may seem like a fair comparison on the surface it doesn’t reflect real life because if someone wanted a 20 ILD comfort layer that’s what they would choose regardless of the density which would mean that different densities with similar ILD’s should be compared to reflect real life choices. The cell walls of lower density Talalay would be thinner. The testing parameters are also different from the testing described on the Latexco site and different types of testing may be more favorable than others to certain materials. They also talk about impact loss for all ILD’s “on average” which doesn’t deal with the specifics of the materials that would actually be used in different layers. Averages don’t differentiate between softer and firmer materials and can make the softer materials seem more durable than they really are.
All of this means that each new set of datapoints needs to be “translated” to some degree … sometimes with some reasonable or educated guesswork or extrapolation involved … so that different sets of information can be more accurately compared.
With this type of technical and complex information (and if you get into some of the math it becomes even more complex yet) … the more you know the more you realize how little you know and how much more you may want to know before you really believe you know anything at all. At least that usually seems to reflect my own ongoing experience and learning curve 
Phoenix