If you seek the highest average damage, yes. However, that is a really common fail approach.
[...] Average DPR is too simplistic to maximize efficiency - it is an overutilized crutch that results in a lot of misperceptions on the power of SS and GWM.
I disagree. I think it's an essential first step. Simply put, if you don't know what your expected damage output even is, you'll find it essentially impossible to draw any further conclusions. It's limited to only look at expected damage, but it's certainly not useless. It's a safe, if naive, assumption that you want to do as much damage as quickly as possible in combat. If you don't know how to do that, you can't really decide whether it's best to use -5/+10 at all.
Yes, if you want to account for overkills, you can use additional knowledge to account for that. But
before the game session even begins you can determine that AC inflection point to be aware of in order to do the highest amount of damage. If your enemy's AC is over that threshold,
you don't even have to consider using -5/+10 unless you're literally throwing a hail mary. You can determine it absent any information about your enemy at all. In most combats, you'll learn AC first, then hp. In many combats, you'll be able to determine AC
visually ("He's wearing chain mail and carrying a shield.") whereas in some combats you won't learn hp until combat is over. Sure, you might be facing a swarm of goblins for the thirtieth time and you know they have less than 10 hit points, but those aren't really the kind of combats we're worried about.
I'd also argue that literally tracking opponent hp (i.e., actually writing down how much damage your party is dealing to each opponent each attack -- not that I'm suggesting you're doing this) approaches metagaming, so I'd expect most tables would allow you to only count hp in your head, as it were. Even then, if you've got an entire party attacking and doing variable damage to varied opponents with varied hp, that gets difficult quickly. Point is, you're often going to find yourself in a position where you don't know enough about hp to use that information. Again, you'll often fall back on dealing as much damage as you can as quickly as you can: expected damage output.
Optimize Fun 2.) The player in question likes the big hit, so uses it inefficienctly from a DPR perspective to increase his fun. I played a Dwarf Barby that *always* used it.
While that's true, it's not really useful. "A player might choose to ignore analysis," is quite irrelevant to the validity of the analysis itself. The game does allow you to make tactically or mathematically poor decisions. That's not a weakness of
the analysis, nor is it something that the analysis needs to consider or account for. It's like suggesting that I shouldn't try to understand what a balanced diet requires because I can always choose to eat nothing but cake.