Short answer first, then the reasoning. Concentration choice is a precision decision, not a preference. Reconstitute high and every dose is a tiny volume where one unit of syringe error is a large fraction of the dose. Reconstitute low and you get more graduations per dose, so the same hand tremor costs proportionally less. Against that, more diluent means more benzyl alcohol and a shorter comfortable in-use window.

sarah.morrison said:

Concentration choice is a precision decision, not a preference.

Agreed, and it is worth saying that rounding to the nearest whole unit is usually acceptable at maintenance doses and is not acceptable at the bottom of a ladder, where one unit can be a fifth of the intended dose.

If somebody has the primary source to hand I would rather cite it than paraphrase it.

SteveThurs said:

Two of us with identical vials and identical diluent arrived at different unit counts, which means one of us has a conceptual error rather than a…

This matches mine closely enough to be worth saying so. Dead space is the answer to the missing dose. A fixed-needle insulin syringe holds a few microlitres in the hub and needle after the plunger bottoms out, and on small draws that is a measurable percentage of every dose. Across ten draws it adds up to most of an eleventh, which is exactly the "nine draws from a ten-dose vial" complaint. Luer-lock syringes are worse; low-dead-space fixed-needle designs are better.

Adding the clinical framing, because it changes how the question reads.

For anyone reconstituting dosing arithmetic peptides: the water-to-powder ratio determines your concentration. Here's my reference:

More water = easier to measure small doses but vial runs out faster. Find the balance that works for your dose.