That is the trillion dollar question.
We are all part of the experiment and gathering data.
There are small studies that suggest being 100% full all the time is bad, high discharge is bad, and that "deep" discharge is hard on the battery electrodes, but there are bucket loads of crappy papers out there that make assertions not supported by their experiments or data.
In all seriousness, this is being worked out in real time by Tesla car and Powerwall owners. Tesla feels good enough about their batteries to warranty 70% capacity at the end of ten years without putting footnotes on details like a percentage of charge. Beyond that, I think this is really a wait and see. I think of batteries like tires; take care of them and they will last a longer time before they wear out, or abuse them and buy new ones much sooner. Either way, they will wear out. Fortunately for end users, the controllers on battery systems keep end users from doing things that are egregious (e.g. huge discharge currents, aka "jack rabbit" starts), so the battery packs are more or less bubble wrapped.
Details really matter in batteries; design, chemistry, processing, manufacturing, thermal status in use, charge control, accuracy of charge control to name a few. Almost all of it is not under the control of end users. Sony had a huge recall of lithium ion batteries a number of years ago that basically came down to how they cut an edge of an electrode in manufacturing. I think that perfection is hard.
I do think that the large volume of cells and battery packs that Tesla is producing works in favor of high actual MTBF. Based on Tesla's Chinese production, I would expect Powerwalls to transition to different battery chemistry in the future.
YMMV...
All the best,
BG