Jason laid it out nicely. I’ll add just a little. Not sure about some of the Chinese cell manufacturers but with the big brands the cells are charged fully and then discharged to nominal voltage (3.7v for most….3.6 for some older chems/models) for storage and shipping. They do this for quality assurance and then for best-practice care while the batteries are dormant waiting to be installed and used by product manufacturers (or us shunned end users with single-cell off label uses). I don’t think any cell retailers really do further testing/charging but there are at least a couple that do check arrival voltages and may perform voltage checks while they have the stock and/or before shipping them out to customers. Not sure if this or anything else happens with manufacturers that are assembling packs or including cells in other devices.
That 30% state-of-charge figure is indeed required for safer shipping (in the US and many other countries although there is some slight variation, depending on country and device/capacity/type). That brings it roughly to the 3.7 nominal voltage. That said, I have received cells that were above 3.8v and many that were around 3.4v or a tad higher, and twice received some that were above 4v.
There are two lines of thinking here. For transport, it’s all about reducing risk, and cells that are at a low state of charge have much less chance of significant venting or arcing if they happen to get mashed or shorted or something…not risk-free of course, but less. The other is just the known best-practice for retaining maximum capacity and maximum cell cycle life. This comes from years of exhaustive testing by the big brands as well as observations from users. Basically when we charge up a li-ion to 4.2v, it’s considered 100% charged but it’s actually an overcharge which is not ideal for the ideals. For us as flashlight users, it really isn’t a big deal and most of use don’t worry about charging them there (most chargers are programmed to do it anyway) and don’t worry about storing them at home at that full charge voltage. If you reduce the charge termination voltage you lose a touch of capacity/run time, which is fine for most lights but will usually affect turbo modes a little, but in doing so you increase cycle life of the cell significantly. As for storage, you can consider the (minimal) self-discharge rate of the cell and also the potential to permanently lose a little of its capacity - the higher the state of charge, the higher the discharge at rest and the higher the loss of capacity. This is also highly dependent upon temperature (room temperature-ish is standard, if you can store them a bit cooler that’s better but no need to refrigerate them (which would also introduce humidity concerns for most fridges)).
So if you really want the best, don’t charge them up quite as much, drain them to around 3.8v give or take a little for storage, and keep them in a nice museum environment with comfy temperature and low/normal humidity. Or mostly ignore it and just run with it…generally fine. The one thing you do want to avoid if possible is extended storage at higher temperatures, where you will see some permanent degradation, and try not to discharge them too low all the time (3.2v to 3.4v is functionally empty but of course they’ll still run on low modes/low current draw for much longer…3v is a good low limit but if you go lower that’s not horrible as long as you don’t allow them to remain at those low states for a long time…just charge them up reasonably soon and they’ll be fine. And discharging below 2.7v…increasing risk of damage but they usually recover fine if you charge them.).
A lot of this best-of is much more important to the primary manufacturer of the cells and to manufacturers of devices that buy many thousands of cells at a time for production use. Everyone wants to deliver their customers grade A performance and to minimize defects/returns which are costly. And it’s perhaps much more important when we are talking about packs which use many cells together in various configurations. But for flashlights, especially just single cell models, not a huge deal unless you just want to adopt the methods.
Here are a couple of links to BU. Great site, quite exhaustive and not always easy to navigate, and some of it is “old” information but largely as true and relevant today as it was when written. Easy to get lost in the sauce and technical info here, but these are a couple basic articles (many more if you feel like going down the rabbit hole):