I feel like the real issues thread like these tend to run into is the focus on Damage Per Round to begin with. I feel like it's very similar to the early days of tracking professional basketball statistics.
DPR is basically Points Per Game. It's the most simple, basic level of evaluation for an individual character/player. It does not take into account a massive amount of the game. In basketball, it would ignore how often you miss, rebound, steal, assist, block, and even more advanced concepts like your value over replacement player, reduced offense from the person you're guarding, wins produced while you are on the court, and all sorts of elements.
And for years those more "advanced" statistics were ignored in basketball because they were more difficult to obtain data on, and more difficult to interpret. Everyone knew what Points Per Game meant and how to get that number (you look at the score board) but very few knew how to obtain further information or how to read that data.
It seems to me D&D is sort of still stuck in that early statistical read on the efficiency of a character. If your goal is "experience points" or "treasure" or "survival" then "casting charm person on the chieftain of a tribe which avoids combat and gets the party XP and treasure similar to if they would have gotten if they had killed the entire tribe" should have high value. But if DPR is the only thing you assess, it has zero value unless your charm person spell caused extra damage to be dealt.
Just because it's hard to come up with some more advanced assessments of efficiency doesn't mean it shouldn't be attempted. Isn't it time we had some stats that measured "avoided getting hit with X amount of damage avoided" or "prevented harm by avoiding an encounter with X amount of draining of resources" or "reduced the risk of death of a PC"?