The one thing about grading overhealing...
Depending on your party (and what content you are running) it might be part of the overall strategy to get a successful run.
If you have a party full of really good players this is probably not going to be the strategy. However if you are in a party full of people who seem to have terrible gear, or not know how to dodge, or try but still fail because they have high ping or whatever, and you're a healer who decides not to just quit right there and then, that group might need things that lead to some overhealing.
That big aoe that good players dodge (but this particular party can't?) Maybe a pre-casted medica ii might be a good idea, it'll do some overheal but the regen will help once half the party decides to stand in the bad. Same with Succor. This leads to overhealing at first, but is done in anticipation.
The increased healing in such parties would also mean less dps because of the increased need for mana preservation and because of the constant need to make up for other people's errors (aka Larrizaur's "Healers adjust").
Another example was a POTD run I had the pleasure to watch (literally. I was watching someone else's monitor). The healer kept waiting until about 25% left on the tank to heal. Normally this is ok, but in POTD, where hidden traps might exist and some of the monsters do a lot of damage for a 4-man party to handle, that was a recipe for disaster.
Instead of anticipating the fact that the tank was taking almost 35% damage per hit from the mobs on this particular floor and topping off (which is a general no-no in high end content because of the overhealing that often comes with it), several times the tank died because the healer was playing "correctly" rather than playing to the party's needs at that time. In this run a "topping" strategy, even with some overhealing, would have led to a more successful run overall.
The formula would therefore also need to account for a lower DPS output when more healing has been done.
So sometimes you have to play "incorrectly" to increase a party's chances of winning depending on the party composition and how they are playing, relying less on meta and more on knowing the actual skills and when they are useful. This would be unbelievably difficult to grade.
A similar dilemma would exist for tanks (or dps that pick up more slack by rezzing, mana shifting, and other support things).
I'm not against the idea but this would be really hard to make it truly balanced.