more urgent for older models as their cross-overs are likely to be needed (due to age)
Good point! I see that as a secondary benefit. The intention of the upgrades is to notably increase the quality of the passive parts and, thereby, the SQ. As Tom Thiel wrote earlier in this thread:
Thiel was always about finding the optimum point on the cost-performance slope so that real music lovers could afford our products.
Speaking for myself (and not Mr. Thiel), the upgrades seek to maximize SQ beyond what was possible at the price points of the original products. IMO, a cost-no-object approach to the XOs for the CS7.2 or CS3.7 would have likely have placed those speakers on par with the best available. The upgrades intend to seek to squeeze the ultimate SQ from the designs.
I am still burning in my new resistors, so don’t yet have a final verdict but my initial impression is that the Mills MRAs are a worthwhile, yet subtle, improvement over the ERSE resistors originally installed. This change in parts would have added $200-300 retail to the CS2.4 and represents, I think, the kind of decision Jim Thiel would have made when optimizing the “cost-performance slope”.
In my case, I recognize the drivers in my CS2.4s are really frackin’ good. IMO, I would have to move up to something like the Vandersteen carbon or TAD beryllium drivers to find something clearly better. Speakers with these drivers are an order of magnitude more expensive than my Thiels! But I think I can get really close to their level of performance by optimizing the XOs. In other words, it's a solution that gets me close to the very best but at a price I can afford.
Also, keep in mind that the technology has advanced since Jim Thiel designed the original XOs. Even the relatively recent Clarity SA coax feeds in my CS2.4SEs have been notably surpassed by Clarity CSA.