When I speak about purpose I tend to speak about the typical usage case. Of course there are other ways to use ultimates but I usually cite what is closest to the developer intention.
The “High of PoTG” is the sensation caused by getting an exceptional amount of on-fire or a long sequence of kills/actions. You ever see a Genji try and 1v4 a team with his blade? It’s because they’re chasing that. The reason why people try and do things like that is because if they somehow manage to pull it off, they can say they carried or otherwise. Unfortunately, this tends to have a low success percentage in most cases.
That could be another proposed change, but if that was the case, it would have to be potentially higher than Lucio’s due to how tide-turning that can be.
As for what is defined as “success” there is two elements to this. Did you finish the casting animation, and did you actually use this to your advantage? You can finish the animation for Visor for example, but it doesn’t matter if you get denied by the shield. With Mercy, the only goal is to get in range and press the button, then there’s nothing the enemy can do (barring a change to the design). If you resurrect even two people, the ultimate is a success. The problem with the old ultimate was due to how much potential it had, people wanted to be “more successful.” Not to say that this isn’t an issue with other ultimates, but with Mercy the issue was prevalent due to the conditions needed to get maximum value out of it.
I would argue that the design is not bad. Uninspired, sure. But the actual utility of being able to damage boost your entire team for 20 seconds is scary. It’s not even like Supercharger where it can be broken in mere seconds, or is made pointless if your team goes out of sight, since Mercy can move to meet them in relatively safety in the sky.
As for failing to get value, it gets value in every situation it would go off. It can get “less”, but not “no value.” Think about shatter, if shatter misses, literally nothing happens. Same with Tracer stick, Grav, Blade, etc. Even Area of Denials like Mei, Hammond, and soon Torbjorn can fail due to either deflect, defense matrix, dying during the animation, or simply not being in the area. The only criteria for success for an ultimate is “doing what the ability is meant to do.” If Mercy resurrects even one person, she has succeeded in doing her job. It might not have been good, but she did it. And due to the nature of the invulnerability it had, that’s why it was frustrating to play against.
The problem with you saying that it “just resets the fight” is that this is simply not true. You don’t get refunded the ults you expended, while the enemy has all of theirs, as well as their ammo refilled. Simply put, it severely punishes using more than one ultimate to get kills due to how easy reversion is.
As for you, I don’t go on the forums a large amount. So I was not aware of you, or your proposed changes. I do think that the ultimate could potentially be worked out, but it would require changes that you listed, mainly being the LoS and lack of invincibility. The reason why Mercy res as an ability is not as offensive in its current state is due to the reasonable counterplay that can happen. There is a significant risk to Mercy upon casting. The ability would need to make Mercy incredibly vulnerable and have a significant charge and cast time, due to the game state changing nature. From a personal standpoint though, I don’t think it would be very enjoyable.
From my own personal standpoint, for about my first 40 hours playing Overwatch I played nothing but Mercy at all. I was not particularly good at FPS games, having not played a PC shooter at all prior to this one. So I valued that Mercy was designed as an easily accessible character that could provide value. The issue with this is that typically, when a character is designed to be easy pick up, they should plateau in value. Events like Moth meta should never happen because it discourages usage of more difficult to use characters with mechanics that provide varied gameplay. The reason I prefer Valkyrie is that from a balance perspective, it makes more sense and provides significantly more counterplay and gameplay style variety. In Valk, you can DPS, you can boost, you can heal, you can mix those things together. Even if it’s only an extension of her kit, it still adds a “superman” factor that resurrect simply does not. Resurrect felt very sloppy and counter to the premise of teamplay, as well as narrow in gameplay style. It was hard to tell the difference between one Mercy player to another at a higher level due to lack of that variety.
Do I think Mercy is perfect in her current state? Not at all. But I think that reimplementation of an ability that not only narrows playstyle but also struggles to keep the balance for the ult economy is bad for the game as a whole. I would rather they give Mercy another option and promote more individuality from Mercy player to Mercy player.