Thank you for the word on the plastic nubs, Max. While mine do not show the problems that Original Poster described Y-E-T, I will be sure to try your fix should mine show them. I do not run mine much, any more. In fact, I might try to sell them on Fee Pay.
I suspect more the presence of the idler truck, which lowers the factor of adhesion, than I would the difference in the metal. This is due to my experience with the RS-1 and RS-3, as well as the Kato RS-2 and RSC-2. I find Atlas' B-1-1-B gearing arrangement curious, but I am guessing that it was easier or cheaper to gear it in that manner rather than A-1-A, although there are plenty of A-1-A geared models out there. In fact, I do wonder about the cost argument, since I recall paying the same for the Kato RS-2 as for the RSC-2, despite the RSC-2 model's being geared C-C (the prototype was, of course, A-1-A). I did not pay list for any of the foregoing, so I do not know if the list prices between the four and six axle models were different.
My experience with the RSD-4 vs. either the RS-1 or RS-3 was not dissimilar from other posters to this topic. The B-B geared models pulled better than did the
B-1-1-B geared model RSD-4. One thing that I did note, though, was that when I ran a pair of Atlas RSD-4s, the pulling power difference between them and a pair of the B-B geared models was only one or two cars. The paired B-B geared models did pull better, but only slightly.
I had only one Atlas/Kato RSD-4, which did have the C-C geared trucks of the prototype. I traded it away long past, so I never had the opportunity to try to put those trucks onto the Chinese frame.
The Kato RSC-2 does pull slightly better than does the RS-2.
As far as I know, Atlas has been silent on its reasons for gearing the model RSD-4 as it did.