According to Agner when the uops go back to to the reorder buffer to get retired they are still treated as fused. However, if you look at the descriptions of the events [1, 2] a retired fused uop is counted as 2 retired uops.

And yes, if you do mov r, m / add r, 1 / mov m, r instead of add m, 1 no such fusion happens. This is specific to CISCy instructions that do more than one thing at the same time.

[1] https://software.intel.com/sites/products/documentation/docl...

[2] https://software.intel.com/sites/products/documentation/docl...