Some very thoughtful comments, and it looks as though you have really tried to develop a workaround.
Here is mine:
My suggestion is not limited to your specific problem. For example, I have discovered that for certain mallets, in pitched percussion, VSL has reduced the range of playable notes on the instrument -- you can write the notes on the staff in Sibelius, but they won't play, because the samples aren't there. It's as though VSL took a saw to a piano keyboard. So, to get the notes I want, I have to choose a mallet I don't want. That problem, too, can be corrected as I suggest here.
As I am working up a composition in Sibelius 7.5.1, I do my best to make it sound as I want it to. I still get the benefit of analyzing compositional structure from looking at a score, but I don't worry about how it looks too much. When I stop working and make an audio or video for submission, I use that somewhat bizzare looking audio copy, after making adjustments for mixture and loudness -- using speakers, not headphones. This copy will also contain text notes, such as identifying a motive, I have written on the score to help me see the structure I have created. I then save this as a separate audio, or "play" copy. This copy makes the audio for video for submission.
I then save another "print" copy. This is supposed to be the professionally correct, playable version of the score. In this I clean up the appearance of the score, changing dotted notes to tied ones so that players can see the beats, for example, checking multiple stops on strings, changing clefs for certain instruments if necessary, general work up of appearance or crowding of notes and staves, adding cues for parts, and other things I can't think of right now. However, it could be at this stage that you would renotate the areas where you want ties to be in your print score. As you point out, your score speaks to players, while Sibelius speaks to a machine. You just cheat on Sibelius a little bit, not playing the "print" version to create your audio demo.
I should add that in the final clean up and polishing, I sometimes (actually, ALWAYS!) find musical problems that need correction, so I have to return to the "audio" version to adjust that as well. And then you have to have a good memory so both scores, as you make adjustments, are the same. So wait as long as possible before undertaking this dual-score approach.
Not perfect, but it seems to work pretty well.