Agreed, that the site I checked with isn't accurate. But it is a site run by a fully authorized Honda/Acura dealer, though, and everything in their catalog is genuine Honda/Acura stock.

I also obviously agree with you that the staff at this dealer site messed up the SKU transcription. But I continue to maintain that Honda's part numbering system is a problem here. A part that looks similar but isn't functionally compatible needs to have not only a different part number, but a visibly much different part number.

And other non-part-number codes should not be shared across similar-but-incompatible parts. Both the GSR and the Type R belt say "126RU26" on them, in the largest lettering on the belt. Easy to make a mistake.

The LS/RS belt says "125RU26". That reflects good practice. Greatly reduced risk of a mixup with a GSR 126RU26 unit.

If they had carried over that nomenclature and changed the leading numeric, I would have looked at my old GSR belt, which says "126RU26" and then looked at the R belt saying "127RU26", and that would have been much more obvious. And I probably wouldn't have wasted any time at all on this matter.

As noted, other people are obviously experiencing the same confusion as I did (including probably some parts desks in the dealer network).

As far as ebay goes, I've bought lots of parts there and have only rarely had bad experiences. I have never been burned by any seller with 100% positive feedback and over 50 sales. People who have worked that hard to build a good record are pretty protective of it.

Anyway, following up on the trick to artificially tighten the belt by pulling on the tensioner spring during the tensioning procedure -- it seems to have worked as intended. Engine has been running fine and making good power from idle to redline. I'm still mildly nervous about pushing this hack hard, and I will probably put a proper GSR belt in the engine in a few weeks, but for now the car runs happily.