I don't understand your post.
I proposed a possible solution. I agree it's not working correctly. SE have had several years to work this out and it seems this is the best they can get. So instead of trying to further optimise netcode they may just have to stick a plaster it. I wonder if the QA tests in house with 0 connection lag can dodge moving out at 99% cast and if they test from JP to the NA datacentre.
I'm British, I eat sleep and breathe sarcasm...Maybe it's just too early for me and not enough coffee ><;;