Hmm. So when you talk about "mitigating" the -5 part of the -5/+10 things, how are you contrasting this with "what you'd get if you used those same tactics to gain plusses, and didn't offset them with the -5 to hit"?
Say you have a tactic which allows you to get +5 to hit, precisely offsetting the -5. Is taking the -5/+10 necessarily better than keeping the +5?
If you're not taking hit probability into account, then you are definitely getting bogus data, MHO.
If the hit rate is changing (as in toggling -5 to hit on and off, etc) then absolutely.
That said, the higher the base bonus to hit, the less impact on DPR the -5 has. In your post 667, you figured a base bonus to hit of +6 (you also seem to have compared +11 to hit with +1, that is, factoring in the "mitigation" in the first case but not the second).
So let's recalculate that first:
Say you have a base to-hit bonus of +6, your base damage is average 5 points with a +3 damage modifier, you have a -5/+10 ability, and you have the ability to "mitigate" the -5, which is to say, you have the ability to add +5 to hit somehow.
You now have three options: +6 to hit, average damage 8. +1 to hit, average damage 18. And +11 to hit, average damage 8, using the +5 without the -5 to offset it. The +6/8 only beats +1/18 if AC is 19 or 20, and breaks even at AC 18. But +11/+8 beats +1/18 from 15 and up, ties at 14, and is only worse against 11-13.
The second option should be +6, 18 (ie the -5 penalty has been "mitigated"). What you've shown is that the is a range of AC (zombies and the like) where even -10 to hit for +10 to damage would be a good trade off!
Anyway, +6 to hit for 18 damage beats +11, 8 at AC 22 and below: vs AC 22, the +11, 8 damage attacker hits half the time, for DPR of 4; the +6, 18 attacker hits one-quarter of the time for DPR of 4.5.
AC less than 23 are most ACs in the MM (certainly vs 1st level PCs).
Now let's increase the PCs' level a bit, by changing the baseline to +10 (+4 prof, +5 stat, +1 magic), and see what happens.
With a "mitigation" of +5 to hit available, options become; +15 to hit, 11 damage (5 +stat + magic); +10 to hit, 21 damage (with -5/+10).
Roughly speaking, the first option will beat the second if the -5 penalty halves the chance to hit, which is to say (roughly) if 11+ is needed to hit, which is to say (roughly) that AC is 26 or better. Vs ACs of 25, the +15, 11 attacker will hit 11/20, for DPR just over 6; the +10, 21 attacker will hit 3/10, for just a bit more over 6. (Note that, though the bonus to hit has stepped up by 4, the optimisation threshold has stepped up only by 3, because the base damage has grown and so +10 is a proportionately small boost.)
ACs less than 26 are most of the ACs in the MM even vs high level PCs.
In other words, if you can "mitigate" the -5 by getting a +5 to hit, and your main concern is DPR, you should use your -5/+10. (Yes, there may be issues with overkill; or a particular reason that you have to hit
now - but I don't think those semi-edge cases detract from the main picture.)
Advantage has a much larger proportionate effect on success chances when you need higher rolls to hit, and a -5 to-hit penalty means higher target numbers on your dice. Advantage is a 50% increase in chances when you need an 11 or better; it's a 95% increase in chances when you need a 20.
Just to spell this out a bit - if you need a 20 to hit, you will miss 19 in 20 time, and hence hit 1 in 20 time. With advantage, you will miss 19^2/400 times, with is to say 361/400 - so you will hit 39/400 times, or near-enough to 1 in 10. That's very close to a doubling of the chance to hit, but I don't think it's very helpful for thinking about typical DPR calcuations in modern D&D, because the design of 4e and 5e is such that it is very rare for the players to need to roll 20 to hit - to hit numbers down around 12 or so (eg 1st level characters trying to hit AC 17 and 18) are far more typical.
A -5 penalty changes a to-hit number of 12 to one of 17, which takes to hit rates from 9 in 20 to 4 in 20; that is to say, a bit less than half. Provided +10 to hit is a bit more than doubling the damage (which it is if base damage is 8: 18/8 = 9/4, which is the reciprocal of 4/9) then nothing has been lost (on average). Now at 1st level I think worries about overkill, and about killing this thing
now (ie at least a moderately steep discounting for next-round kills, when some PCs have hp in the single digits), are pretty salient - and so +5/-10 isn't that attractive. But it is more likely to come on line at mid-levels anyway, and I think by then it does get more appealing, especialy as to hit bonuses might have grown by +3 (stat, prof, magic item) while ACs, due to bounded accuracy, may be more static.
Between 3e's Power Attack and 5e, the designers decided that getting to choose the penalty/bonus for each attack wasn't as fun/balanced for some reason. I think it was 4e that moved to the current rule, and 5e just stuck with it figuring that it works better for most games.
4e's Power Attack is -2 to hit, +2 per tier to damage. I've never seen it used - I haven't checked the maths in any detail, but my gut feel is that (i) hitting in 4e is probably more important than 5e, because you're more likely to be carrying non-damage riders on your hit, even as a non-caster; (ii) the characters who get the hit bonuses that would mitigate the -2 (eg rogues with daggers) have sufficiently big damage totals that the damage bonus is not proportionately significant enough.
Eg if your hit chance is, by default, 13 in 20 then -2 drops it to 11/20. So the bonus to damage has to be at least a two-elevenths increase to balance out. But most heroic tier PCs will do more than 11 points of damage: eg a rogue might do 2.5 from dagger, 7 from adding two stats, +7 from sneak attack; a fighter might do 5 from weapon, 4 from strength, 3 from other buffs; etc. There are non-DPR-focused epic builds that do less than 33 on average, but they are probably doing other things on a hit which mean they don't want to reduce their hit chance so badly.
In other words, I think the 4e feat is not that good.
I'm done with this thread, folks.
That would be a pity.
I agree that [MENTION=12731]CapnZapp[/MENTION] is, to some extent, missing the wood for the trees, but it's still an interesting thread with some worthwhile ideas being tossed around.
Rather than looking at nerfing the -5/+10 (and thereby nerfing fighter damage into irrelevance) the focus could be on making melee more viable - this also has the benefit of helping make the non-GW/SS options for fighter viable (eg if melee is more of a thing, than melee-range battlefield control becomes relatively more important).
Maybe closing rules need to be looked at. Or even doubling the length of a round, and hence doubling movement rates, but leaving everything else (including rates and ranges of attacks) unchanged?