Why I Have Yet to Not Despise Mercy's Current State


You all know why I’m here by this point. This is my 5th thread on this topic, and as you can see, I’m not limited in thread title variations. Great news, though! I got bored of my previous text on top of the fact that I had lots of time to think while on vacation, so I’m writing this entire thing from scratch!


Let’s dive straight into this.

Balance! Or… Lack Thereof:

Resurrect is currently the reason Mercy is imbalanced.

How so, you might ask, citing Mercy’s Resurrect statistics from before her rework and now, which are very close in numerical value?

Ignoring the process of elimination that would lead us to Resurrect in the first place, the difference lies within how Resurrect is used.

First is the availability of Resurrect now versus Mercy’s 1.x versions. Today, Resurrect runs on a consistent 30 second cooldown regardless as to the Mercy player’s performance. Killing a Mercy repeatedly does not impede upon their ability to return and use Resurrect; the only thing Mercy needs to have Resurrect available is time, and nothing in the game can deprive her of that. Mercy 1.x, on the other hand, needed to fill a resource meter in order to use Resurrect. Sure, that resource charged passively over time, but at an incredibly slow rate proportional to the requirement. If the player were to just wait for that resource to fill passively, they would be waiting 5 minutes and 25 seconds, which could be half the match by itself.

Anyway, because Mercy needed to fill a resource in order to use Resurrect on an acceptable basis, it was far easier to impede its availability by targeting or eliminating the Mercy. The time respawning and walking back to the fight would otherwise be time spent healing, shooting, or amplifying damage, depriving her of ultimate charge and therefore a use of resurrect.

Second is the optimal usage for the ability. Given Resurrect’s current availability and single-target mechanics, its optimal usage is obvious; it is best used to reverse the first pick in a fight, operating as a 1-up for an ally in every poke-at-choke scenario. This one-up functions as a buffer that disincentivizes an enemy engagement after the enemy gets a pick and therefore a numbers advantage, resulting in a continuation of the poke battle and a second chance for Mercy’s team to get the pick instead. As a result, Resurrect now stops the snowball before it begins, making it much more reliable when it comes to post-rez success. Preventing a snowball is much easier than trying to reverse a snowball after it has already gained velocity.

Mass-Resurrect, on the other hand, had variation in its optimal time of usage due to its inconsistent availability and its variation in numerical value. If it were to be used as a tempo-rez every time, it would simply be outmatched by the post-rework Resurrect because it cannot compete with 2.x’s cooldown rate. Thus, in order to maintain a good numerical value, a balance between tempo and mass-revives needed to be held. Finding that balance and by extension the optimal time and placement for Resurrect was difficult, making the ability harder to use overall.

As mentioned before, however, it is more difficult to reverse a snowball after it has gained velocity. Thus, Mercy 1.x’s Resurrect was far less reliable in post-rez success than Mercy’s 2.x’s Resurrect.

Resurrect is also used much earlier in the fight than it was before, reducing the window to prevent it from what used to be anywhere between the first pick and after the fifth kill to just the first pick. There is a smaller window to stop it.

As though the above advantages weren’t enough to push Resurrect over the edge, Resurrect also no longer occupies one of the team’s six ultimates. An ultimate is not lost upon its expense.

To recap:

  • Availability is consistent and unaffected by performance. It is easier to have resurrect ready to be used now than it was with 1.x.
  • Optimal usage is predefined, making it much easier to use Resurrect to its maximum capacity now than with 1.x.
  • Now prevents snowballs before they begin, which is much easier than stopping them after they are already in motion. Much more likely to end in success.
  • Is now more difficult to prevent through preliminary action, as it is used much earlier now.
  • No longer takes up an ultimate slot.

While the number of revived players every match may be nearly identical, the mechanics of the newer Resurrect makes it far more powerful in application.

So what does this mean in regards to Mercy’s balance state?

Resurrect will always be difficult to balance as a basic ability. Either it can be feasibly used with or against a given composition, in which case it’s too good to pass up; or it cannot be feasibly used with or against the composition, in which case it simply is not worth considering. This is why Mercy’s GM pickrate post-rework has consistently been above 12% or below 2%. It is next to impossible to find a decent balance state for Mercy when her most powerful ability is a basic ability.


When the rework first went live, there were already a lot of restrictions on Resurrect. It had the longest ability cooldown in the game by far, a 5 meter range, and a line-of-sight requirement.

And yet, Resurrect, and by extension Mercy, was absurdly overpowered.

Mercy’s ultimate, Valkyrie, provided additional uses of Resurrect; three extra uses, to be exact. In the first post-rework patch to Mercy, that number was dropped to 1. On top of this, Guardian Angel’s cooldown no longer reset upon Resurrect’s activation, making hit-and-run revives more dangerous.

The developers came back to see that once again, Resurrect, and by extension Mercy, was absurdly overpowered.

In response, the developers implemented a 1.75 second cast time and a 75% movement speed reduction to Resurrect. To top it off, Resurrect could now be interrupted not only by CC, but by knockbacks as well. If interrupted, Resurrect would begin its 30 second cooldown.

The developers returned to see that yet again, Resurrect, and by extension Mercy, was absurdly overpowered.

To fix this, the develop- Okay, hold up.

How the hell did the developers not figure out the issue by this time? Were they not paying attention to their own actions?

You took a mess of an ability with a metric [REDACTED]-ton of exceptions, and then you applied even more exceptions to that ability. How have you not seen the issue yet? Have you ever stopped to think about why you need to apply these exceptions, and what these exceptions resemble?

Fast forward to today, and there isn’t much left in Resurrect that can be reasonably nerfed given its current position… and yet, it is still insanely powerful, to the point where you are nerfing other areas of her kit to make room for it. Have you every stopped to think about why that might be?

Have you ever considered why those exceptions are exceptions?

What abilities in the game have a downtime comparable to a 30 second cooldown?

What abilities in the game have a cast time longer than one second, and movement penalties and ability disables beyond that?


Resurrect is an ultimate. Look around at the other abilities in the game. The only ones with comparable downtimes, cast times, and cast penalties are ultimates.

  • Resurrect - 1.75 second cast time. 75% movement speed reduction, inability to use attacks and abilities. Ability activates after cast time ends.
  • Configuration: Tank - 1.5 second cast time. Inability to move, use attacks, or use abilities. Ability activates after cast time ends.
  • Rocket Barrage - 3 second cast time. Frozen movement (including vertical movement). Ability is active during cast time.
  • Earthshatter - .6 second cast time (+any time it takes to hit the ground). Ability activates after cast time ends.
  • Dragonblade - 1 second cast time. Inability to use attacks and abilities while casting. Ability activates after cast time ends.
  • Tactical Visor - 1.4 second cast time. Inability to use attacks and abilities while casting. Ability activates after cast time ends.
  • Sound Barrier - .8 second cast time (typically, can be reduced by moving onto a higher surface. Can also be increased by jumping beforehand or moving onto a lower surface). Inability to use attacks and abilities while casting. Ability activates after cast time ends.

When you need to add self-inflicted penalties onto an ability that match or exceed the penalties imposed by ultimates, that’s a pretty clear sign that said ability is an ultimate. Thus, it should not be a basic ability.

Even Jeff Kaplan has outright stated this issue:

“We tried to move Resurrection to a secondary ability, and the ability, right now, in current Overwatch, isn’t playing out as a secondary ability; it’s playing out like another ultimate ability”

January 25th, 2018.


6:10 - 6:23.

This is why Resurrect as a basic ability will not work. Trying to force an ultimate into a basic ability is suicide not only from a balance perspective, but also from a gameplay perspective.

Obviously, a cast time and self-incapacitation on a basic ability doesn’t feel good. What happens to Mercy after she presses E is no longer under her control; it is in everyone else’s.

It doesn’t feel good to wager your own fate against the incompetence of your allies and the coordination of the enemy. It doesn’t feel good to lose control of your own character.

I am expecting this to be brushed off by people who supposedly don’t care about how a hero feels, but I might point out several things:

  • This is a videogame. The player is supposed to feel empowered by their actions and their character. Turning into a rock for two seconds doesn’t meet that criteria.
  • If it was your hero(s), you would probably feel and respond the same way a lot of Mercy players are now.
  • Mercy was reworked in the first place in part because “she didn’t feel good to play against”.

The self-inflicted penalties on Resurrect render the ability not as an asset the player would jump at the opportunity to use, but as a chore they player must endure, as using the ability is objectively better for the team’s chance of victory than not. This just so happens to be the reason Symmetra’s original Photon Barrier (the ORIGINAL one, the one that was part of her kit before her first rework) was removed.

By a show of hands, who likes to be incapacitated for nearly two seconds because you need to cover for the likely-stupid mistake(s) of an ally?


Didn’t think so.

Cooldown Versus Charged:

When Resurrect was an ultimate, it, like all ultimates, needed to be charged through the player’s actions. As it was often one of two or fewer support ultimates designed to protect or recover the team from four or five offensive ultimates, Resurrect was a resource that was scarce in supply compared to its demand. Combine this with the fact that this ultimate was also the only reliable high-impact asset in Mercy’s kit at the time, and the Mercy player’s success depended heavily upon their ability to quickly charge Resurrect. Minimizing the downtime of Mercy’s ultimate was a must.

This glaring need to have Resurrect primed became most obvious in the moments leading up to a teamfight. If the Mercy did not already have Resurrect, the pre-teamfight phase was generally the time in which they would rush to gain those last few points. Aside from teamfights themselves, this pre-teamfight stage was typically when ultimate charge was at its most plentiful for both teams, and the Mercy needed to have their ultimate up prior to the teamfight’s commencement.

During these pre-engagement scenarios, both teams typically face off against one another at a choke point or from opposing sides of a divide created by the map. It is often referred to as a “poke battle”, as that does a good job of describing what the two teams are doing in relation to each other. Each team is scanning for weaknesses in the enemy to capitalize on. Both teams are looking for a fight, but they also want to draw a decisive win out of it; they want some insurance that the fight will go their way rather than the enemy’s. That insurance usually means an advantage upon engagement… the most common of these being a numbers advantage from an early kill.

As a result, both teams are poking each other, testing each other, looking for that weak spot to press on and secure a numbers advantage through a kill. At the same time, they want to prevent the enemy from doing the same to them; each team is minimizing the window of opportunity for the enemy to hit them. Consequently, damage is not often inflicted in large quantities, but when it is, it is dealt in very high quantities over a very short period of time. Heroes that move slightly out of position are quickly moved even further out of position, whether that’s because they are forcefully moved towards the enemy or the enemy suddenly moving towards them.

Playing normally at this time would not provide the Mercy player with enough ultimate charge to have Resurrect ready when they need it; damage isn’t dealt in a steady stream at this time, unlike during teamfights. In order to charge Resurrect, the player is forced to play more actively. Rather than sticking with the bulk of the team behind barriers, they might fly to and heal that Genji off to the right, who is attacking the enemy from the high ground… then they might fly to that Soldier:76 behind the team and amplify his damage for a few shots… then they might fly back to the tanks in the frontline.

If no opportunities for gaining more ultimate charge presented themselves, the Mercy player would create their own opportunities by drawing their pistol and laying down some of her own fire.

Mercy’s playstyle went from reserved, defensive, and attentive those in secure positions, to a more aggressive, almost hungry, mobility-based playstyle. Healing allies that were out of position was not something that was done purely for the purpose of keeping them alive; in many cases, attempting to sustain them was a lost cause. Instead, the Mercy healed them because healing them provided more ultimate charge than not.

Normally, this way of playing would be considered too risky. Flying out in the open away from the rest of the team was already a big risk for the hero with the highest target priority in the game. Flying to a single ally who was already drawing the enemy’s attention and fire only added to that danger. Shooting at the enemy with her pistol means being able to be shot at.

Trying to charge Resurrect quickly during this time meant jeopardizing one’s safety. However, the alternative was to enter the next fight with yet another risk; not having Resurrect. The Mercy is given a choice:

  • Play defensively now and take few risks, but have one big risk to face later: A greater chance of losing the incoming teamfight.
  • Mitigate that big risk by taking smaller, more manageable risks now: Play more aggressively.

Between the two choices, the latter one is typically better, due to its risks being more under the Mercy’s direct control; it is easier to control the outcome of a situation that largely revolves around oneself and a few other players. When the situation includes all twelve players, like the incoming teamfight would, Mercy’s contribution and therefore influence is comparatively insignificant. Having more control over the risks means they are more likely to end in the player’s favor.

However, they are still risks nonetheless, and just like when playing any other hero, taking these risks and coming out on top feels good. Getting in, grabbing some ultimate charge, and getting out alive left the player with a giddy feeling, as though they just got away with something they should not have been able to.

The need to charge Resurrect created its own minigame of balancing risk and reward. Lose the balance in one direction, and Mercy would be missing an incredibly important resource when the teamfight begins. Lose the balance in the other direction, and the Mercy would overstep their boundaries and get killed before the fight begins. Hold the balance, however, and the player is rewarded with a use of Resurrect and a feeling of accomplishment. Enduring this playstyle successfully felt like pulling off a dangerous stunt.

This game of balancing risk and returns not only added diversity to Mercy’s playstyle, but it also gave the player something they could be proud of. If a player charged Resurrect from zero to full in 40 seconds while weaving in and out of enemy fire, they could take pride in the fact that another player probably would not have been able to do the same. It gave the player something to work towards and improve upon, as knowing the hero’s limits, and persistently pushing said hero’s limits without overstepping them, bore rewards.

Today… Well, none of what I’ve said above still applies. Resurrect is now spoon-fed to the player at a constant rate regardless as to how actively or passively they play. Valkyrie fails to fill the void, as it doesn’t have any playmaking or tide-swinging capacity.

The player can no longer take pride in the availability of Resurrect, as there’s nothing to take pride in with that regard. There is no longer a larger upcoming risk that can be compensated for by taking smaller risks now, so there simply isn’t any point in taking unnecessary risks. The diversity in playstyle and pride taken in a successful aggressive endeavor simply isn’t there anymore; those endeavors are now maladaptive.

Before the rework, Mercy players could be differentiated by their playstyle. Good Mercys were visible, yet seemingly untouchable. Inexperienced Mercys were either invisible or far too tangible. Today, there is no difference. Every Mercy from mediocre to absolute-god is a ghost for most of the match.

There isn’t a difference in Resurrect’s availability anymore either. A player could die twice in 30 seconds and still respawn with Resurrect primed.

See that top-500 Mercy that saved her tanks while dodging three ultimates before turning around and quadruple-tapping the enemy’s Tracer? See that Bronze Mercy that just got hooked and killed by the enemy Roadhog at the beginning of the teamfight? They’ll have Resurrect available at the same time. There is no room to grow or improve.

Now we are left with a bland and somehow-infuriating static cooldown on Resurrect.

“No, I don’t have Resurrect now, team, but don’t worry; I’ll have it right after we all respawn.”

Thought Process and Usage Reward:

Remember when the cast time was placed on Resurrect and how a lot of players responded with “Now Mercy mains will have to think before using Resurrect”?

Do you also remember that this is how I introduced this section the past two times I made this thread? I’m bringing this up again because I thoroughly enjoy irony.

Anyway, there is a very clear difference between the complexity of using Resurrect now, and the complexity of using Resurrect prior to the rework. Stemming from that difference in complexity, there is a big difference in the thought processes for the two abilities… namely that ult-rez had one, and E-rez doesn’t.

Before delving into the thought processes, I’ll go over the differences in complexity.

Currently, Resurrect has a restrictive numerical value, allowing for only six possible combinations of Resurrect targets:

  • Solo-rezzing allies 1, 2, 3, 4, or 5. 5 combinations.
  • Not using Resurrect. 1 combination.

Ult-rez, on the other hand, had far more possible combinations of Resurrect targets:

  • Solo-rezzing allies 1, 2, 3, 4, or 5. 5 combinations.
  • Duo-rezzing allies 1 through 5. 10 combinations (5 choose 2, for you math freaks out there).
  • Trio-rezzing allies 1 through 5. 10 combinations.
  • Mass-rezzing (four-man) allies 1 through 5. 5 combinations.
  • Team-rezzing all allies. 1 combination.
  • Not using Resurrect. 1 combination.

That is thirty-two possible combinations of Resurrect from Mercy 1.x, making it more than five times as complex as the current iteration of Resurrect, and that’s from a numerical perspective alone.

Resurrect currently has a 5 meter radius, a 1.75 second cast time, and a 75% movement speed reduction, narrowing the parameters in which it can be safely used. It is less flexible due to safety concerns.

Resurrect used to have a 15 meter radius, no cast time, and post-rez invulnerability, allowing it to be used in a variety of situations. It was very flexible.

Resurrect currently is not an ultimate ability. It is on a 30 second cooldown, making its availability predictable and reliable. On top of this, its limited power range does not allow it to contest other ultimates.

Resurrect used to be an ultimate ability. It had no guarantee as to when it would be available again, making it more difficult to gauge when it could next be used. It was not only an ultimate, but a support ultimate, rendering it a scarce resource. It was capable of matching the power of other ultimates, and its expense counted as an ultimate expense.

As a result of lower numerical complexity, usage inflexibility, downtime predictability, and a basic-ability status, E-Resurrect also has a consistent optimal use scenario. That optimal time of use is always to reverse the first pick prior to a teamfight. This is the case for several reasons…

First, there is safety. With Resurrect’s incapacitation upon cast, it is clear that Mercy could use all the help and protection she can get while casting; she already has the biggest target in the game on her head. The more living allies nearby to protect the Mercy in the first place, the more likely it is that using Resurrect will result in a successful Resurrection and a living Mercy. Using the ability when there are only two allies to protect the Mercy is typically suicide.

Second, the realistic impact of Resurrect is greater the earlier it is used. Its numerical value remains the same (one person), but that numerical value means much more when used earlier in the fight rather than later. This is because of how teamfights typically snowball.

Both teams start out facing each other at a dividing point in the map. Six players on each side. One team gets a pick (or a first kill that grants an advantage and a prompt to engage), and then pushes into the opposing team. The team of five players is at a one-man disadvantage, a weakness their enemies capitalize on with their engagement. The team of five loses another player, dropping them down to four. Then they lose another. Somewhere in there, they may have gotten a kill, bringing to fight to a 3v5, but it doesn’t matter. The team that got the first pick has too much momentum and too great of a numbers advantage for the losing team to fight back with any success. The team of now three players crumbles, and the last three living players are killed off.

What point in that fight would be the best time to use Resurrect?

Using Resurrect when it’s a 3v5 or worse would probably mean not being able to get Resurrect off in the first place… but let’s cast that aside (no pun intended) and say that the player somehow manages it anyway. Okay, great! You got that ally back, and chances are that another ally is in critical condition from the damage they sustained over the past 1.75 seconds of you not healing them. That is, assuming they didn’t just die in that time.

But you know what? Let’s ignore that second flaw too. Let’s suppose that miraculously, Mercy pulled off the Resurrection and every other ally that was still alive was left unscathed during that 1.75 second cast time… It’s still a 4v5. Mercy’s team is still at a huge disadvantage.

Okay, let’s try using Resurrect earlier… How about on the first pick?

When trying to revive that first dead ally, four other teammates are there to cover for the Mercy; she has a good chance of getting Resurrect off, so that’s a good start.

What about Mercy’s other allies? Well, the enemy doesn’t have much momentum yet and there are four allies to spread damage among rather than only two, greatly reducing the chance of more allied deaths or even the chance of them dropping to critical condition. Both flaws presented in the late-fight Resurrection are heavily mitigated when using the ability earlier.

Not only that, but a successful Resurrection means the fight is back to a 6v6; the enemy no longer has the numbers advantage, and has likely backed off because of that. Teams typically want to engage with an advantage.

It is easier to stop a snowball before it starts than it is to try to stop it once it is rolling. Rather than using Resurrect to little or no effect later in the fight, it is best to prevent the scales from tipping in the enemy’s favor in the first place. Therefore, using Resurrect as early into the fight as possible is always the best option.

Resurrect during Mercy’s 1.x versions, however, was a different story. It had greater numerical complexity, it was more flexible in how it was used because it had no activation safety restrictions, its next use was unpredictable, and it was contending with other ultimates. As a result, it never had a consistent optimal execution. Its best-case scenario varied from fight to fight.

For example, if the Mercy waits to use Resurrect on four or more players every time, she’s reviving her team into an uphill battle every time; the enemy has the momentum, the positioning advantage, and the first shot. Furthermore, waiting to revive a bunch of players at once runs the risk of staggered deaths, risking a numbers disadvantage on top of those other disadvantages.

On the opposite end of the spectrum, using Resurrect to revive 1-2 players every time ran the risk of blowing Resurrect without depleting enemy ultimates first. Considering that “Helden sterben nicht!” was practically a green light for the enemy team to start activating ultimates, using Resurrect early in the fight against a loaded team without a plan B wasn’t really a good idea.

How was the best execution of the old Resurrect determined? That is where the difference in thought process between the two ability versions begins. The optimal time of use is predetermined for the current Resurrect, while the optimal use of the old Resurrect needed to be discovered on the fly for each individual scenario.

As for how the best execution of old Resurrect was determined… It depended upon a lot of variables. For starters, here are some variables and questions considered before the engagement begins:

  • Living allies prior to the teamfight. Are we a full team, or are some allies dead/walking back from spawn?
  • Living enemies prior to the teamfight. Does this put us at a numbers advantage/disadvantage? Is it possible that additional enemies will join the fight late?
  • Ultimates held by present allies. Are there any allies I should prioritize when it comes time to Resurrect because of their ultimates?
  • Ultimates held by present enemies. How many ultimates does the enemy team have at the ready? Does this put us at an advantage/disadvantage in primed ultimates? Are there any ultimates I should be weary of in particular?
  • Missing enemies. Are there any enemies I cannot see/have not accounted for? Are they flanking? Where might they attack from? Am I their target?
  • Allied positions. When things get hot, who can I fly to for safety? Will I still be able to heal the frontline from there, or will healing the tanks require another reposition?
  • Mercy’s position. How can I place myself in a way that still allows me to support the team, but minimizes the enemy’s ability to shoot at me? If I am directly attacked, could a reach a position that gives me an advantage in that fight? Are there nearby healthpacks that I could take advantage of?
  • Engagement patterns. Will the enemy get a pick and then engage, or will they fly through that choke point, ultimates blazing? If we are the team initiating, which pattern will we follow?

During the engagement, here are some of the variables the player might consider in the heat of battle:

  • Midfight numbers. Are we at an advantage/disadvantage in this category? Will Resurrect give us a numbers advantage? Is Resurrect even necessary to win the fight?
  • Enemy activity. How aggressive are they playing? Are they slowly gaining an upper hand, or has all hell broken loose? How aggressive is each one playing? Do the aggressive ones have their ultimates ready? Are they prepared for me to use Resurrect, or are they too blinded by their tunnel vision to realize I’m still here?
  • Death locations. Are my allies close enough to one another to be revived in a single use of Resurrect? Is there a specific position I should attempt to reach to revive them all? Would reviving a particular ally provide our team with an advantage due to their location in relation to the enemy?
  • Midfight ultimates. What ultimates are active? What ultimates do we still have? What ultimates does the enemy still have? How can I bait out those ultimates? What ultimates do our dead allies have? Are they in a position to use them?
  • Respawn timers. Will more allies die before that first ally respawns? Should I revive that first ally now, or should I hold onto resurrect until more allies are down?
  • Resurrect priorities. If I am forced to choose between reviving one ally or the other, which do I choose, based upon their hero, skill, and ultimate status?
  • Enemy positions. Where is each living enemy? Are any flanking around in an attempt to kill me? Am I in the sights of any enemies?
  • Safety/practicality limits. Is it too dangerous for me to have a presence in the fight anymore? Would healing my team even be useful at this point? Is it time for me to pull back and supplement healing with pistol fire?
  • Post-rez outcome. Given all of the above variables, what are the chances that my team will win the fight after I resurrect them? If the chances are not in our favor, how can I tip them in our favor? Who should I try to finish off with my pistol while taking advantage of the invulnerability frames? If the post-rez fight will not be in our favor, should I bother using Resurrect at all?

There was a mind game that centralized around defining the best way to use Resurrect. It was a mental puzzle in a race against time. The mind game brought an additional layer of player engagement to a mechanically basic hero.

In contrast, Resurrect currently has a static optimal execution. The only real thought that goes into the ability’s use is safety, and its optimal execution serves to mitigate its hazards anyway. Rather than a series of questions, hypotheticals, and planning that require close attention to the fight, quick thinking, and refined awareness skills, Resurrect now only requires a check for a few specific parameters based upon the positions of visible allies and enemies.

“Will doing this get me killed?”

If the answer to that question is “no”, then it is clear to use Resurrect.

Analogy time! Who’s ready for an analogy that I totally haven’t used before?

Go outside and stand in front of (what I am assuming is) your friendly suburban neighborhood street. Now cross it, but don’t be an idiot about it. Be safe. Do what your parents drilled into you twelve-thousand times: Look both ways before crossing the street.

Look left. Are there any cars coming?

Now look right. Are there any cars coming?

If the answer to both of these questions is “no”, you may cross the street. Otherwise, you should wait for those cars to pass before crossing. Simple, isn’t it? The most difficult part of this entire “mental exercise” is remembering to check both ways in the first place, and that isn’t a concern anyway; it’s common sense.

Are you able to do this on a regular basis? Yes? Congratulations! You have mastered the art of not getting yourself killed while using Mercy 2.x’s Resurrect. The train of thought and mental demand between the two scenarios are the same.

Resurrect’s complexity was lost through the rework, removing its mind game and a layer of engagement. Resurrect hardly requires any thought to use anymore, and there isn’t any variation in how it should be used.

Players used to claim and complain that Mercy was a Resurrect bot… Mercy became more of a Resurrect bot because of the rework. Resurrect has been reduced to nothing more than a bot-like function.

Resurrect from 1.x felt very good to use. It presented the player with a dilemma that, if worked through and executed correctly, left the player feeling heroic. They knew that they changed the course of the fight through their own actions. This should be the ultimate goal of a videogame: To make the players feel like they contributed something valuable because of something they did, because of a challenge they overcame. To present the player with the opportunity to do something special by the terms of their own performance, making the task difficult, but allowing the player to take full ownership and pride for a job well done.

Resurrect from Mercy 1.x was all of the above for the player. They earned Resurrect by charging it in the first place, they examined the battlefield, they planned ahead, they evaded the enemy threats, and finally, they flew in to save the day.

“Heroes never die!”

Resurrect’s dramatic effect only added to the player’s experience. It was a buildup of tension followed by a heroic release. The animation and voice line were very appropriate for the ability.

A tide-changing play that the player can take fully ownership over, further enhanced by the mechanics and cosmetics of the ability used to make that very play. These are the moments that make the player truly feel like a hero, and this nature in Resurrect is why the ability became so iconic in the first place. It made the player feel like a hero.

Now… that feeling is gone. It has been gone for a long time.

Resurrect is given to the player regardless of performance. It feels less like something that the player worked hard to earn, and more like something they received as a participation award or as a crutch. Once there is a prompt to use Resurrect, the player takes a lazy glance at the situation and decides whether or not it is safe to use Resurrect. If it is, they fly in, endure nearly 2 seconds of Stone Simulator, and then reverse some random person’s early-fight pick for the third time that match. The only thing a player can really take ownership for in that situation is having the tolerance to play Mercy in the first place. There is nothing special about a use of Resurrect. If one player can use Resurrect in a given situation, chances are that any other player in their place could do the same.

A rewarding and iconic ability that created a layer of engagement on its own was replaced by a bland shadow of itself that feels average at best to use.

How the developers are so blind that they cannot see this is beyond me.


Why is this thing still an ultimate ability?


Valkyrie is a disgusting… pile? I’m trying to think of a good word to describe it…

Looks up synonyms for “pile”


Valkyrie is a disgusting conglomerate of existing abilities that were all thrown into a single ability with seemingly no regard for cohesion or player experience. There’s nothing original in Valkyrie.

Flight? Pharah has that.
Mediocre AoE healing? Hello, Moira, Lucio, and Ana.
AoE damage amplifier? Orisa.

As a result of trying to put all of these into one ability, the developers also made every ultimate that Valkyrie contends with (or tries to, anyway) simply outclass it. The ability can be summarized by “Jack of all trades, remotely sufficient in none”.

Valkyrie poses the opposite problem that Resurrect does. Rather than an ultimate placed behind a basic ability, Valkyrie is a basic ability placed behind an ultimate.

Let’s list what Valkyrie currently offers:

  • Allows Mercy to fly in any direction at 9 meters/second. (5.5 is base walking speed for most heroes).
  • Extends the range of Mercy’s beams from 15 meters to 30 meters.
  • Extends the range of Guardian Angel from 30 meters to 60 meters.
  • Mercy’s passive regeneration is no longer interrupted by damage.
  • Mercy’s beams spread to all allies within 10 meters of the primary target.
  • Guardian Angel’s flight speed is increased.
  • Mercy’s pistol has infinite ammunition and a faster projectile speed.

If we were to cut this list down to the parts that actually have a direct influence on the fight, we would get:

  • Mercy’s beams spread to all allies within 10 meters of the primary target.
  • Mercy’s pistol has infinite ammunition and a faster projectile speed.

That’s it. Everything else either has no realistic use, or it just makes Mercy harder to kill… which doesn’t really matter being that she removes herself from the fight while using Valkyrie anyway.

Let’s throw these capabilities up against the ultimate abilities that are most comparable to Valkyrie: Transcendence and Supercharger.

Transcendence is very similar to the healing aspect of Valkyrie in several ways. Both heroes gain additional mobility upon activation. They both heal all five allies in a 10 meter radius around Zenyatta/the primary beam target. They both become (effectively) unkillable while using their respective ultimates. They are both unable to impact the fight in any way other than healing while healing their team with their ultimates; Zenyatta is a channeled ultimate, and Mercy is unable to heal and shoot at the same time.

What’s the difference between Transcendence and Valkyrie? Transcendence has five times the healing output (soon to be six!). Transcendence heals for 300 health/second on every target. Valkyrie heals for 60 health/second on every target. Valkyrie’s healing rate is simply blown out of the water by Transcendence.

Unfortunately, we do not have statistics that tell us exactly how much healing Mercy is doing through Valkyrie, but we do know that Mercy’s overall healing output really isn’t much higher than it was pre-rework. As of the day I am typing this (August 7th), Mercy’s average healing output is 12085. Her pre-rework healing numbers were around 11900. If we were to just assume that Valkyrie’s introduction is the only thing responsible for the current statistic, we would only be able to say that Valkyrie provides a bonus 200 healing/game on average. In contrast, Zenyatta averages 2259 Transcendence healing/game.

Supercharger is one of three ultimates that increase a player’s damage output, the other ones being Valkyrie and Nano-Boost. Its visuals are similar to the visuals of Valkyrie’s damage amplification beam, and both ultimates amplify the damage of all allies in an area around the Supercharger/primary beam target.

The similarities stop there, and just about every difference between the two abilities favors Supercharger over Valkyrie. For starters, Supercharger has fire-and-forget mechanics. Orisa can deploy it and resume playing as usual, as she does not need to channel the damage beams for them to be active. In contrast, Mercy becomes the supercharger, effectively removing the team’s main healer from the fight while she channels her damage beams. Furthermore, not only does Orisa get to participate in the fight while her ultimate is active, but she also receives the damage benefits from her own damage amplification. Valkyrie is five players and one buff dispenser; Supercharger is six players and one buff dispenser.

As if another source of damage was not enough for Supercharger to outclass Valkyrie’s damage amplifier, Supercharger also increases the damage of all affected targets by 50%, instead of Valkyrie’s 30%. On top of this, it also has a substantially larger AoE radius than Valkyrie; 25 meters. The one downside Supercharger has in comparison to Valkyrie is that it needs to be protected, but considering the rest of Orisa’s kit, that isn’t terribly difficult. She has her barrier to cover it and her hitbox is a good size to body-block enemies from attacking the drum, especially with Fortify to help her do so. The only good options there are to destroy the Supercharger are to flank or to get in close… neither of which are great ideas when the entire enemy team is dealing 150% of their normal damage output.

So… Valkyrie is totally outclassed in both the healing and damage amplification categories by other ultimates. Not only does the ability not have anything unique to its name, but it is simply worse in every category than its ultimate counterparts.

Valkyrie is really just a basic ability. Compare its impact to that of other supports’ basic abilities and that becomes clear:

  • Amp It Up (healing aura): 46.8 health/second to all allies within 10 meters of Lucio. Lucio can shoot while this ability is active.
  • Biotic Grenade: An instant 100 health when used on allies, along with a lingering healing amplification. When used on enemies, it deals 60 damage and leaves a lingering status effect that negates all healing received by those affected. Fire-and-forget mechanics. 4 meter blast radius.
  • Orb of Discord: Increases the damage received by one enemy target by 30%. Fire-and-forget mechanics, allowing Zenyatta to both deal damage and heal while his Discord Orb is active. (Virtually) Endless duration and can be retargeted at will when provided line of sight and a 30 meter proximity. 3-second decay time when target leaves range and/or vision.

Give Valkyrie a 3 second duration and slap a 12 second cooldown on it (beginning after the duration ends), and you have a nice basic ability.

But that’s not what the developers wanted. Instead, for reasons never provided, they wanted to take the pile of dung that they scrapped in the alpha stages of the game (citation) and slap it into the game as Mercy’s ultimate. How did they do this?

They extended its duration 20 seconds.

Of course, probably seeing that the ability was still flaming garbage, they decided to add a few other bonuses to it. They added a bunch of buffs to Mercy’s pistol (most of which were removed before going live), and they added the potential to get four uses of Resurrect out of Valkyrie (rez, valk, rez, wait, rez, wait, rez). This time they were successful in making Valkyrie worthy of being an ultimate, because they added an ultimate (Resurrect) to a basic ability (Valkyrie). A basic ability in addition to an ultimate nets an ultimate ability.

Now the developers had a different problem. Mercy had two ultimates: Resurrect and Valkyrie… Or to be more direct about it, Resurrect and then more Resurrect. As you probably already know, Valkyrie was nerfed repeatedly. Resurrect was slowly withdrawn from the sack of garbage that was sacked and then unsacked, eventually rendering the sack of garbage to be a sack of nothing but garbage.

When Resurrect was pulled out Valkyrie, so was the ultimate in Valkyrie. Now we’re just left with a basic ability that lasts fifteen seconds on Q, while Mercy’s real ultimate is on her E.

Configuration: Spectator:

Consider the risk and reward for playing at different proximities from the fight for most heroes. Generally, the closer a hero is to the fight, the more impact they have; targets are bigger, and more enemies are visible. Projectile speed, spread, damage falloff, and limited range on abilities become less of an issue as a hero gets closer to the fight. The closer a hero is to the fight, the more they have to shoot at/heal/protect. By the same token, being closer to the fight also means greater risk, as the player becomes a bigger target for the enemy.

With the exception of a few specific heroes (Widowmaker) the risk and reward scale at about the same rate with proximity to the fight.

And then there’s Valkyrie.

The risk still scales with Mercy’s proximity to the fight while using Valkyrie; the closer Mercy is to the fight, the bigger of a target she is.

The reward does not scale with Mercy’s proximity to the fight while in Valkyrie; her team is receiving the same 60 (or 50, whatever) health/second regardless as to how close Mercy is.

As Mercy gets closer to her team while in Valkyrie, the risk increases. The reward is flatlined. This then prompts the question: Why would any Mercy with even the slightest hint of common sense want to be anywhere near the fight?

Trick question. They wouldn’t. The reason to get in close in the first place is to maximize reward. If that reward vanishes and the heightened risk stays, then the obvious choice is to minimize that risk by staying as far away as possible.

On top of this, Mercy has free flight with an omnidirectional movement speed of 9 meters/second and a 30 meter beam range. Not only does the Mercy have the incentive to be as far from the fight as possible, but they also have the tools to do so with ease.

The result? The best way to play Mercy while using Valkyrie is to hide in the skybox for fifteen seconds while sitting 29 meters from the beam target. This poses very little risk and sacrifices no reward.

This would be fine if Mercy had something to do while floating halfway across the map from the fight, but Valkyrie more than takes care of that. All of the skills and nuances of Mercy’s base kit that keep the player active and engaged? Yeah, those are thrown out the window.

  • Healing prioritization? AoE beams remove the need for that.
  • Damage boost utilization? AoE beams also negate the need for that.
  • Positioning? There aren’t really any threats to evade in the skybox, and you are already in the best position you are going to be in for the duration of the ability.
  • GA discipline? You’re still in the skybox and you have free flight, remember?
  • Awareness? There isn’t much of a need to pay close attention to the fight. It’s not like you would get any real benefit from doing so. Skybox and AoE beams are still a thing.

Consequently, Valkyrie treats the Mercy less like a player, and more like a game object; like a construct. It doesn’t give the player anything to do. Instead, it does everything for the player.

Are you 29 meters from the fight and holding down LMB/RMB? Congratulations! You have reached maximum capacity. That is all you will be doing for the next 14 seconds.

Naturally, an ultimate that is a borderline AFK button doesn’t feel good to use. Whereas most ultimates offer a unique change in playstyle that still demands player activity (Dragonblade, Tactical Visor), offer a brief burst of power (Earthshatter, Rocket Barrage), or leave the hero’s playstyle unchanged (Supercharger, Photon Barrier), Valkyrie transforms the player into a glorified spectator that applies flat bonuses to their team. From the player’s perspective, they aren’t helping their team; they are idle. Instead, Valkyrie is doing all of the work.

And this is a failure in game design. The player and their character’s kit are not merging as one. The player can’t take ownership for the ability and say that they helped their team; they didn’t put any effort into doing it. It was just given to them.

Participation awards and idle-execution abilities isolate the player from their character. It changes the perception from “I did this” to “This was given to me”. The player doesn’t take ownership for the actions that were supposedly theirs, and thus, they don’t feel heroic. Instead, they feel like a sidekick or a tool used to let others take the glory.

It gets worse. Not only is Valkyrie a glorified spectator with an absurdly low skill ceiling, but it isn’t powerful enough to contest other ultimates or even basic focus fire. Valkyrie may have the potential to heal at a combined rate of 300 health/second, but that healing rate is spread out evenly across 5 allies… all of whom either don’t really need it, or need more than 60 health/second to survive. As a result, most of that healing isn’t helpful, resulting in Valkyrie having a nearly identical effect on the fight as a Mercy playing at high capacity without Valkyrie. Hence the “easy-mode” stigma on the ability.

Because of Valkyrie’s inability to sustain a team, not only does the Mercy player feel idle, but they also feel utterly helpless as they watch their team get mowed down/blown up/frozen/knocked down/cut apart.

And this is the biggest problem with Valkyrie. No hero, no ability, no ultimate, should ever leave the player feeling helpless or useless. That is perhaps the worst thing that could ever happen to a playable character, and doesn’t scream anything other than developer incompetence. It contradicts the very point of active videogames in the first place: To empower the player.

Although I guess I should commend Valkyrie for what it is good at. Despite its shortcomings in impact, player engagement, and skill limitations, it is perhaps the most flexible ability in the game. No other ability in the game offers all of these options midfight:

  • Showing the world how much you like the color blue by leaving a weight on Mercy’s alternate fire the entire time.
  • If you are a strange person, showing the world how much you like the color yellow by leaving a weight on Mercy’s primary fire the entire time.
  • Casually shooting down the enemy Pharah.
  • Inspecting the ceiling textures.
  • Fetching a drink from the refrigerator.
  • Drinking some of that drink you pulled out of the refrigerator.
  • Checking your social media (or the Overwatch Forums!).
  • Petting your cat.
  • Petting your dog.
  • Feeding your pet fish.
  • Using a laser pointer to drive your cat nuts.
  • Working on homework.
  • Eating a sandwich.
  • Looking up cat gifs.
  • Reading this entire list.
  • And my personal favorite: Using Valkyrie to fly back to the fight after respawning.

The Recent and Previous Developments...

…From those who are bad at developing.

Recent Developments:

On the PTR, Mercy received another nerf, bringing her number of post-rework nerfs to 13. In order, here is the list of nerfs to Mercy since Mercy 2.0’s introduction:

  • Pistol damage in Valkyrie reduced from 30 to 20.
  • Mercy’s pistol no longer receives an increase in fire rate while in Valkyrie.
  • Valkyrie flight speed reduced from 11 meters/second to 9 meters/second.
  • Guardian Angel’s cooldown no longer resets upon Resurrect’s activation.
  • Valkyrie no longer reduces Resurrect’s cooldown.
  • Resurrect now has a 1.75 second cast time unless Valkyrie is active.
  • Mercy’s movement speed is reduced by 75% while casting Resurrect.
  • Resurrect can now be interrupted through CC, knockback, and death. If interrupted, Resurrect begins its cooldown.
  • Mercy is no longer exempt from Resurrect’s cast time while Valkyrie is active.
  • Valkyrie no longer grants a charge of Resurrect upon activation.
  • Valkyrie’s GA movement speed bonus reduced by 50%.
  • Valkyrie’s duration reduced from 20 seconds to 15 seconds.

And finally:

  • Mercy’s primary beam healing output reduced from 60 health/second to 50 health/second.

That last one is on the PTR now, and based upon the fact that just about every change that hits the PTR goes live, it is safe to assume that this change will also go live. There’s one big problem that I see with this change right off the bat, but because I occasionally like to be suspenseful, I’m going to hold off on that until we glance at the two sets of developer comments for this change:

“Mercy is intended to be able to consistently pump out more healing than any other healer over the course of a match. While this is currently true, the difference in healing is so significant that it makes it very difficult for other healers to compete with her for a spot on a team. Reducing her healing output will close this gap a little, but she will still maintain her status as the go-to pick for raw healing power. We’ll keep an eye on her to make sure she is still a strong pick.”

“Mercy’s previous healing output made her nearly impossible to replace in any team composition. Even after these changes she will still be able to deliver more healing over the course of a match than any other support hero. This change should allow for other healers to be a more viable pick.”

Let’s start picking these apart.

“Mercy is intended to be able to consistently pump out more healing than any other healer over the course of a match. While this is currently true, the difference in healing is so significant that it makes it very difficult for other healers to compete with her for a spot on a team.”

“Mercy’s previous healing output made her nearly impossible to replace in any team composition.”

The first and largest problem here is a misidentification of the problem. Mercy currently is a must pick, but is her healing really the cause of her mandatory status?

I’ll give you a hint: The answer is “No”.

How do I know this? Simple Geometry logic; basic enough to be broken down into a basic truth table using Boolean data types (True/False):

The above table shows the possible relationships between two Boolean variables (P and Q) and the outcome of those two variables joined by an “and” function.

Applying this to our scenario, “True” represents a state of balance; “False” represents a state of unbalance. The variables P and Q represent different parts of Mercy’s kit. “P and Q” represent Mercy’s overall balance.

For a hero to be balanced, the entirety of their kit and each individual part of their kit must be balanced; hence the fact that the outcome of the function “P and Q” is only “True” if P is “True” and Q is also “True”. If one or several parts of a hero’s kit is unbalanced, then the hero is unbalanced.

We know that prior to the rework, Mercy was balanced, almost underpowered in some cases. She had a high overall pickrate, but that pickrate sharply fell off when approaching higher ranks in favor of other healers. She was viable enough to have at least decent representation throughout all ranks, yet she was behind every other healer in GM pickrates, so she could not have been viable to the point of being oppressive. If she was, that would have been reflected in such ranks.

If we know that Mercy 1.x was balanced, then we also know that every part of Mercy 1.x’s kit was balanced; the only factors that will yield a balanced or “True” result are ones that are also balanced or “True” (see the first row in the table).

We know that post-rework Mercy is overpowered, as her pickrates across the board, but especially in GM, are very high. On top of this, the pickrates of other healers in GM (the exception being Zenyatta, who combos well with Mercy) are abysmal, demonstrating just how oppressive Mercy currently is.

How did Mercy go from balanced to overpowered, or from “True” to “False”? Her kit was changed by the rework, and either P or Q was changed to “False”.

The keyword in that last sentence is “changed”. A Boolean cannot go from “True” to “False” or vice versa unless something caused it to do so; something changed in it.

Connect the dots.

  • Prior to the rework, “P and Q” returns “True”, therefore both P and Q individually are “True”.
  • After the rework, “P and Q” returns “False”, therefore P, Q, or both, are “False”.
  • A variable flipping from “True” to “False” can only happen if that variable is altered.
  • The only parts of Mercy’s kit that could have switched from balanced to unbalanced are the parts that were changed by the rework.

Do you know what changed through Mercy’s rework?

Not her healing output.

Her healing output is the same as it was prior to her rework; the same 60 health/second, the same aim-assist allied lock-on, the same ease of use and retargeting.

“But what about Valkyrie?”, I hear some of you furiously fingering into your keyboard.

I don’t know…

Mercy’s average healing/game as of June 28th, 2017 (pre-rework): 11905.
Mercy’s average healing/game as of August 12th, 2017 (pre-rework): 11912.
Mercy’s average healing/game as of March 21st, 2018 (post-rework): 11926.
Mercy’s average healing/game as of May 31st, 2018 (post-rework): 11603.
Mercy’s average healing/game as of June 30th, 2018 (post-rework): 11752.
Mercy’s average healing/game as of July 22nd, 2018 (post-rework): 12132.

What about Valkyrie?

Mercy’s healing output since her latest patch has remained within about 300 HP of where it was prior to her rework.

That said, I’m going to humor this remark anyway. It helps to establish a point.

Suppose Valkyrie was problematic in regards to balance. What should be changed as a result?

I’ll give you a hint: Not anything that isn’t Valkyrie.

If a hero is overpowered, the reason for them being overpowered needs to be defined. Once the issues have been localized, nothing but the source of those problems should be altered. Altering a “True”, regardless as to whether or not it causes that variable to change to “False”, does not change the other variables that are “False” back to “True”. Therefore, the hero will remain unbalanced.

Everything I have said above should come intuitively. If A and B results in balance, but then B is replaced by C, and then A and C results in an imbalance, it’s pretty obvious that C is the source of the problem. The problem will not be fixed by changing A.

Take Hanzo’s rework, for example. Hanzo was underpowered prior to his rework, and his rework, which granted him Storm Arrows and Lunge while leaving the rest of him unchanged, made him overpowered. The only thing the developers should be altering with Hanzo (and the only thing they have altered so far) are his Storm Arrows and his Lunge. Nerfing his primary fire, his Wall Climb, or his Sonic Arrow wouldn’t solve the problem, and would only serve to piss off the Hanzo playerbase.

Sound familiar?

The source of Mercy’s dominance isn’t something that has remained the same for 10 seasons (her healing output). The source of her dominance is something that has changed dramatically since then: Resurrect.

“Reducing her healing output will close this gap a little, but she will still maintain her status as the go-to pick for raw healing power.”

“Even after these changes she will still be able to deliver more healing over the course of a match than any other support hero.”

Really? I’m a bit suspicious. This is a flat 17% reduction to Mercy’s total healing output, and that’s not accounting for the slower charge rate on Valkyrie as well.

Let’s do some math. Suppose that the 17% healing reduction transferred perfectly into her current averages. As stated above, our current average for Mercy’s healing output is 12132.

12132 - (0.17 * 12132) = 10069.56 healing/game.

10069.56 would be our new healing average for Mercy. I wonder what Moira’s is…?

10609… She also received a slight buff to her resource regeneration in the PTR, so that number will likely go up a bit as well. Are you sure that She will still be the best at healing? I don’t think she will be.

…Hang on. I’m getting word that a bug fix has just gone live. Let’s check the patch notes to investigate!

“Fixed a bug that allowed Mercy to damage boost certain abilities (e.g. Hanzo’s Dragonstrike, D.Va’s Self-Destruct, and Junkrat’s Steel Trap)”


Hmm. This “bug fix” reminds me of Lucio’s Wall-Ride “bug fix” in that it has been an intended feature since forever ago and the developers just now decided to remove it but wanted to be stealthy about it.

Seriously. Come on. When amplifying Hanzo’s ult, Mercy gets hitmarkers on her screen and assists in the killfeed. She received (and currently still receives) ultimate charge from amplifying it. Her “damage amplified” statistic increased while boosting these abilities.

Abilities that truly never were intended to be amplified by Mercy’s damage beam (Rip-Tire, for example) did none of the above.

It’s pretty obvious that this has been an intended feature since day one. If you’re going to nerf Damage Boost, at least have the spine to say it rather than trying to kick it under the rug.

In short, my reaction to the recent patch notes can be summarized by this picture:

Shall we move on to previous events?

Initial Reasons for the Rework:

When the Mercy rework was revealed, we were given a developer update that identifies two reasons for the rework.


Timestamp: 0:38 - 2:12.

These two reasons can be summarized by “Hide and rez” and “Resurrect didn’t feel good to play against”. I will start with the latter of those two reasons, as it can be undermined and pulled apart in fewer words.

“Resurrect didn’t feel good to play against”. The first and most obvious question to ask at this point is, “What ultimate, when used to any benefit of the enemy team, is fun to play against?”

Last I checked, being blown up by Rocket Barrage, sliced apart by Dragonblade, frozen by Blizzard, denied by Transcendence, nuked by Self-Destruct, or knocked down by Earthshatter, does not feel good. This is a PvP game; “disheartening to play against”, “frustrating to play against”, or “unfun to play against” are a given for at least one team for part of each match, or the game wouldn’t be empowering for the other team… and thus, wouldn’t be enjoyable. Using this as a reason to change a balanced mechanic isn’t justifiable, unless the same can be said about every other ability in the game. Otherwise, it is simply a double-standard that reveals favoritism.

The second rhetorical question to ask is, “Does it feel better to play against a hero that has been a must-pick for nearly a year?”

Would you prefer to play against a character that resembles every other hero in that their ultimate is annoying to deal with, or would you rather play against a hero that is objectively overpowered, has been nearly a win condition for far too long, and still has an annoying ultimate?

Shall we dive into a bit of player perspective philosophy?

Unless the hero is a balance concern, Player 2’s perception of Player 1’s character isn’t relevant. It is given that Player 2’s perception of Player 1’s character is likely negative when playing against Player 1. On the other hand, Player 1’s perception of Player 1’s character is not a given in any videogame. For the videogame to be successful, Player 1’s perception of their own character, or how it feels to play their own character, needs to be positive. If that perception is negative, players will stop playing. As a result, Player 1’s experience with Player 1’s character is infinitely more important than Player 2’s perception of Player 1’s character.

In short, a player needs to feel empowered by their own actions. The player does not need to feel empowered, and in a PvP game, should not and will not feel empowered, by the actions of their opponents.

Player 1’s experience with Player 1’s character was breached and disregarded when Blizzard tried to make Player 2’s perception of Player 1’s hero positive… which they also failed to do. By attempting to change that which neither needed to be changed nor could be changed, the balance team tampered with something that should never be altered.

Moving onto “Hide and rez”

“We think it’s wrong to tell a main-healing character to go off and hide somewhere and stop healing for some period of time.”

But it is fine for a main-damage character to go off and hide somewhere and stop dealing damage for some period of time, like many heroes currently do in preparation for an ultimate?

In actuality, you are right; just not for the reason you think you are. You are correct because hiding was maladaptive for the team’s chance of victory, and therefore was a bad way to play her. Allow me to explain this in an even more elaborate manner than my past few megaposts.

The success of a mass-Resurrection (4+ revived players) depended upon several factors:

  • The Mercy’s ability to survive and execute the ability in the first place (the obvious one).
  • The number of players alive on each team after the Resurrection.
  • The number of available ultimates held by each team after the Resurrection (but primarily the enemy team).
  • The shock value of Resurrect against the enemy (by far the most important factor for the post-rez fight).

The last factor is more important than the other two post-rez factors because it determines what other factors need to advantage the Mercy in order to have a good chance at a successful Resurrection. As the enemy team is given time to reorganize and fire their first attacks at the newly Resurrected team, they would already have a positioning advantage and the first shot… that is, if the team was expecting the Resurrection.

If the enemy team is expecting the Resurrection, they will have the post-rez advantages of the first shot and additional positioning time. Of course, this is a very bad situation to be in for the newly revived team, requiring some very strong advantages of their own to compensate and have a chance at winning the fight. In this case, they would need both the numbers advantage and the ultimate ability advantage.

If the enemy team isn’t expecting the Resurrection, then the Mercy could probably get by with either the numbers advantage or the ultimate advantage and still have a chance at winning, as the enemy team might not capitalize on their positioning and first shot opportunities as well as they should due to the shock value of Resurrect.

How does “Hide and rez” fit into all of this, you might ask? Hiding before the fight removes the possibility of having enough of those advantages against the enemy team in the post-rez fight, and therefore any chance of a victorious post-rez fight against any team of similar skill.

Hiding leaves the initial fight to a 5v6, the missing hero on the Mercy’s team being the main healer and primary source of sustain. Obviously, this is a huge advantage for the team of 6 in the first fight, allowing them to steamroll over Mercy’s team with likely zero casualties, and ultimate expenditures used only to counter any of the opposition’s ultimates. On top of this, the team of 6 also now collectively has 2000-3000 more ultimate charge from killing 5 players in addition to any healing dealt over the duration of that fight. Thus, it hands the ultimate ability advantage to the team opposing the Mercy.

The opposing team, assuming they are paying even the slightest attention to what is happening around them, would notice that the enemy is crumbling very easily for a supposed 6v6, raising suspicions as to whether or not the fight really is a 6v6, quickly leading to the realization that the enemy Mercy is missing from the fight. From this, only one conclusion could logically be drawn; she’s waiting to use Resurrect. There goes the ultimate’s shock value.

At this point, even if the team of 6 somehow managed to sustain 2 casualties in the initial fight, it won’t matter. If both teams entered the fight with the same number of ultimates, the team that does not have the hiding Mercy has at least two more ultimates than the other team (2000+ ultimate charge, Mercy’s team used Resurrect). One or two offensive ultimates is the absolute maximum a team needs to wipe a freshly-revived team when they are expecting the Resurrection, resulting in the revived team staring down the barrel of a gun they have no chance of avoiding.

In contrast, a Mercy who stays with her team keeps the fight at a 6v6, granting her team more sustain and forcing the enemy to expend ultimates to overpower them. Once the ultimates started firing, the mercy would distance themselves from the fight, evade, or take cover, laying down pistol fire or amplifying damage rather than attempting to heal those for whom healing would be useless. As a result, the enemy was less likely to realize that Mercy has Resurrect (the enemy was comparatively difficult to kill, removing the prompt to think about why they were crumbling easily in the first place), Mercy’s team was more likely to pick off a few enemies, and Mercy has successfully baited more ultimates from the enemy team. As a result, her team has more advantages and fewer disadvantages in the post-rez fight.

The only cost to staying with the team was an increased chance of being picked off before being able to revive the team, which just follows the same risk choice explained in the Resurrect section; if taking more controllable risks now helps to mitigate the high risk later, then taking those smaller risks now is safer than not in the long run.

Consequently, reviving a team after hiding typically resulted in a second wipe, while staying with the team until it was unsafe to do so was more likely to yield favorable results. That is why it was “wrong to tell a main-healing character to go off and hide somewhere”. It was a bad tactic, as it usually backfired.

Rather than letting the issue resolve itself by fixing the SR exploit that ultimately led to the tactic’s rise in the first place, the developers decided to remove Resurrect as an ultimate entirely. They changed the thing that wasn’t the problem, and ended up pissing off the playerbase.

Sound familiar?

Anyway, I’ve used this example many times before, but removing Resurrect as an ultimate because some players abandoned their team leading to a second wipe is equivalent to removing Rocket Barrage as an ultimate because Pharah players too often blew themselves up with it. There’s no need to intervene; the tactic will die out on its own, so long as the ranking system isn’t feeding it.

Ironically, the rework ended up only replacing one tactic of hiding with another. This time, however, the is no drawback to it, making the tactic beneficial. Valkyrie is best used by pressing Q and then flying the hell away, barely staying within beam range of the team. Doing this minimizes risk, but compromises no rewards. A Mercy playing closer to her team in Valkyrie is not helping her team any more than a Mercy 29 meters from her team, and is only exposing herself to enemy fire. Consequently, staying as far as possible from the fight for Valkyrie’s full duration is the way to go.

But wait! There’s more! Jeff Kaplan was interviewed by Venturebeat at one point in which he revealed another reason for the rework after it was pushed live. In the interview, Jeff Kaplan said this:

“The goal — I’m about to tell you the goal and it’s going to make us sound like miserable failures. But the goal was to lessen the impact of resurrect overall on the game”

To be fair, I have never seen a truer quote from anyone who has authority in balance changes. It sure as hell does make you seem like miserable failures. How is month #11 of must-pick Mercy going for ya? Are you ready for months #12, 13, and 14? :popcorn:

Post Rework Contradictions and Failures:

When the initial rework dropped into live servers on September 19th, 2017, this set of developer comments came with it:

“While resurrecting downed allies is a core part of Mercy’s gameplay, the way her Ultimate functioned was causing a number of problems. It was frustrating to play against, and it incentivized Mercy players to hide away from important battles, instead of taking part in them. This version turns Resurrect into a single target ability. It’s still an important part of Mercy’s kit, but plays much better for both Mercy players and her enemies. Valkyrie, her new Ultimate, gives her the opportunity for big game-making plays and opens a number of new options for her.”

I’ve already gone over the reasons for the rework, so I’m going to skip over the first part of those comments. For now, just keep these two quotes in mind while I go grab more developer comments. Try not to laugh (or get very angry, either would be a justifiable reaction to these quotes) in the meantime:

“This version turns Resurrect into a single target ability. It’s still an important part of Mercy’s kit, but plays much better for both Mercy players and her enemies.”

“Valkyrie, her new Ultimate, gives her the opportunity for big game-making plays and opens a number of new options for her.”

Jeff Kaplan was interviewed by Venturebeat, who published an article with his answers from said interview on November 3rd. I’m not going to post the entire interview, as that would extend this post by a lot, much of which wouldn’t even be relevant to the topic at hand. Keep this quote in mind too:

“We are very intentionally trying to bring the power level of resurrect down. We want the top two things you notice about Mercy to be, ‘Oh my God, she’s an amazing throughput healer,’ and ‘Wow, look at that mobility, she’s just flowing back and forth through the map.’ And we don’t want her to be the resurrect bot that’s just erasing kills.”

On November 16th, Mercy’s second nerf after the rework went live (the previous one being on October 17th). This was the cast time nerf on Resurrect. You know, the one that drops her movement speed to 25% and disables attacks and abilities. With this patch, we got these developer comments:

“We feel that Mercy’s recent rework has been successful, but her Resurrect ability still feels too strong and frustrating to play against. Now that it has a cast time, enemies are more able to counter the ability. However, Resurrect will cast instantly when Valkyrie is active. This should make her feel powerful when she transforms on the battlefield.”

Keep holding those laughs. Or anger. Or whatever emotion you are currently suppressing. Let’s connect some dots… or rather, show where the developers failed to do so.

“This version turns Resurrect into a single target ability. It’s still an important part of Mercy’s kit, but plays much better for both Mercy players and her enemies.”

Three and a half months later, Resurrect had a 1.75 second incapacitation tied to it with no exceptions.

It’s pretty intuitive that an ability that disables a player for nearly 2 seconds is going to play much worse for the player than an ability that doesn’t. Not being able to control your character in a game where being able to control your character is a given does an excellent job of sabotaging the gameplay.

So yeah… I would trade the current rez for mass-rez without a second thought. Because it’s clear that mass-rez played better for Mercy than the current version of Resurrect.

That is despite her must-pick status. Oh look, a segue into the next part of the quote!

“…plays much better for both Mercy players and her enemies.

I guess I should ask this question again. Does it feel better to play against a hero that has been a must-pick for nearly a year?

How about a full year, because that’s where we’re headed at this point?

Moving on…

“Valkyrie, her new Ultimate, gives her the opportunity for big game-making plays and opens a number of new options for her.”

“This should make her feel powerful when she transforms on the battlefield.”

Those new options, the new opportunities for big game-making plays, and the feeling of being powerful when transforming on the field now consists of sitting in the skybox 30 meters from the fight in a slightly-less-useless-than-spectator mode… Probably as your team dies beneath you to enemy ultimates.

Such power. Many game-making plays. Impressive PotGs. Cool new options.

Well, I guess that last one is true. Now you can do things like fall asleep, feed your cat, or make a sandwich!

“We are very intentionally trying to bring the power level of resurrect down.”

Then how about you do that instead changing her healing, which has been fine for the past 10 seasons? Or are you going to keep banging your head against the wall?

“We want the top two things you notice about Mercy to be, ‘Oh my God, she’s an amazing throughput healer,’ and ‘Wow, look at that mobility, she’s just flowing back and forth through the map.’”

Ten days later, Resurrect’s cast time went live. If you want Mercy to be about mobility, why was your first order of business after saying that to place a standstill mechanic on one of her basic abilities? And again, why are you nerfing her healing when the problem is Resurrect?

What you say and what your patch notes say flatly contradict each other:

Developers: “We think mobility should be one of Mercy’s core mechanics…”
Patch notes: “…So we made Mercy turn into a rock for two seconds while using one of her abilities.”

Developers: “We think healing should be one of Mercy’s core mechanics…”
Patch notes: “…Which is why we are scapegoating and nerfing her healing rather than the ultimate on a basic ability.”

Whatever to make the rework seem successful, I guess?

“And we don’t want her to be the resurrect bot that’s just erasing kills.”


Sorry. Just… enjoying this a bit too much.

You say this when:

  • There is no variation in how Resurrect is used. There is always a single best moment to use it, and that is to reverse early-fight picks. You have removed any thought required for the ability’s perfect utilization beyond a check for single set of parameters, reducing the planning process to an automatic function. A bot-like function, to be less subtle.
  • Resurrect is the reason she is a must-pick right now, and will continue to be a must-pick for another few months, due to your mind-numbingly slow, and yet somehow still rushed “balancing” updates.

Developer/Community Interactions - Misleading, or Outright Lies?:

As you all probably know, we now have a 10th garbage bin megathread garbage bin for Mercy discussion, the endless stream of feedback dating as far back as August of 2017.

All along, we have been receiving these messages:

There was also this from Jeff Kaplan on the old forums:

“Thanks to everyone for taking the time to write up thoughtful feedback here. We’re reading the responses daily.”

Most recently, there was this quote from Tom Powers. It was not necessarily on the topic of Mercy, but still in regards to feedback:

Clearly, they would know that the bulk of our complaints are not in regards to Mercy’s balance state, but how Mercy isn’t fun to play anymore, right?

January 25th, 2018:
6:39 - 6:43

“If we tone her down too much, we will bring her back.”

The developer update is titled “Popular Community Topics”, yet Jeff Kaplan uses all of the time spent on Mercy to talk about only her balance state. He doesn’t even mention how she actually feels to play.

He also said this on the old Overwatch Forums:

"We have no plans to revert Mercy. We also feel like she’s not in a horrible place or unplayable.

It’s still to be determined if she is exactly where she should be because it has been too soon and the dust needs to settle. Her playtime in QP and Comp is still incredibly high. Her winrate is still above 50% but more in line with what we’d expect.

But I’d like to put the notion that Mercy is going to be reverted to rest.

Also, we don’t agree with the statement that all supports are weak. Support heroes are extremely powerful and impactful in OW."

Lots of discussion on Mercy’s balance, and not a single mention as to how Mercy feels to play.

More recently:

Even more discussion on Mercy’s balance state, still no sign that they have a clue as to what our concerns are. They were too busy addressing a nonexistent concern to do that, apparently.

Why might this be?

Allow me to reference someone else who took the time to look at the statistics:

"Reading our Feedback"

"Jeff Kaplan: 5.1k posts read
Geoff Goodman: 6.8k posts read
Bill Warnecke: 3.6k posts read
Romain Dijoux: 1.2k posts read
Scott Mercer: 851 posts read.
Aaron Carter : 33 posts read
Micheal Chu: 14.1k posts read
Tom Powers: 15.7k posts read
[Feedback Thread Continued Part III] Mercy Updates - Jan 30, 2018: 17.7k posts; 100k posts overall.

Exactly who is reading our feedback when the contents of one thread exceeds the read posts of 6 Overwatch Developers?"

So that explains the disconnect, but it raises another question: Were they lying about reading our feedback?

The most obvious and simplest answer is “duh, of course”, but there’s one tiny detail that needs to be identified.

In every post made that is supposed to assure us that someone on Blizzard’s end is reading our feedback, they always say “we”.

Right. So, uh… Who is “we”?

Note that it is never said that “The Balance Team is monitoring your feedback”, or “Developers are monitoring your feedback”. It is always “We are monitoring your feedback”.

On the rare occasions that moderators or developers talk to us, they do so as Blizzard representatives. As a result, “we” can be anyone that works for Blizzard. It could be moderators that operate halfway across the world from the balance team, who have no say in such decisions, and it still qualifies as “we”. The developers could be totally unaware that the Overwatch Forums exist, and still say truthfully, “We read your feedback”.

It is not an outright lie to cover up the fact that they don’t care. It is even more malicious. It is intentionally misleading, designed to tell a lie without actually being one.

So the next time you see a developer or moderator say “We read your feedback”, ask them who “we” really is.

A Solution

Unlike the past four iterations of this thread, I had more resources to draw upon while creating this post. Specifically, I had the ideas and feedback of 45+ other Mercy players. Collectively, we created a 20-page Google Document with new ideas for ultimates, basic abilities, and modifications for Mercy’s passives. We then voted on those ideas through a survey I created to determine the best modifications to Mercy, or to rephrase and be more direct about our objective, the ideal Mercy rework.

Here are the results from such endeavors:


  • Removed.


  • Removed as a basic ability and placed as Mercy’s ultimate ability.
  • 15 meter range.
  • Line of sight requirement. Does not apply to destructibles, payloads, or non-fundamental map obstacles that are too small to actually obscure a player. In other words, none of that BS where a player can move behind a lamp post to evade a nuclear explosion.
  • 1 second cast time. Basic movement is unrestricted while casting, but all attacks and abilities are disabled for Mercy during this time.
  • If Mercy is stunned or killed while casting, Resurrect does not activate and Mercy loses all ultimate charge.
  • All living allies within range and LoS of Mercy once the cast time ends receive a burst heal of 150 HP. Mercy also receives this burst heal.
  • All dead allies within range and LoS of Mercy once the cast time ends are revived at full health and receive 2.25 seconds of invulnerability (same as current Resurrect and Resurrect pre-rework).
  • Mercy does not receive invulnerability upon Resurrect’s activation.
  • Resurrect does not require dead allies to be activated.
  • 1625 ultimate charge requirement (the same as Mercy 1.x’s Resurrect).

Most importantly, this takes Resurrect off one of Mercy’s basic abilities and places it into a position that matches its ultimate-like behavior. On top of this, various trade-offs were made to the ability to allow for the healing burst and the introduction of a second basic ability.

The healing burst, post-rez vulnerability, cast time, and LoS requirements all incentivise using Resurrect to revive 0-3 players rather than 4+ players. Attempting to use it on four or more players runs a greater risk a failed resurrect attempt due to the cast time. If Mercy manages to use Resurrect, there is a greater chance that she missed some allies due to the LoS requirement. If those two drawbacks do not drag her team down enough, there is also the fact that Mercy is vulnerable post-rez while her teammates are still reviving.

The healing burst and fire-at-will mechanics on Resurrect make the ability more flexible and open up some unique uses for it. For example, Mercy could use Resurrect to grant her team a burst of healing in the middle of the teamfight to keep their momentum going rather than waiting for an ally to rejoin the fight after being killed and revived.

New Basic Ability - Pacify:

  • After a .5 second wind-up, Mercy launches a single hitscan attack from her left hand.
  • The projectile’s hitbox size resembles that of Symmetra’s secondary fire.
  • The projectile deals no damage upon impact with an enemy. Instead, it applies a debuff that reduces the base damage dealt by the target by 33%. 100 damage is reduced to 67 damage before other damage amplifiers/reductions are factored in.
  • This debuff lasts for 4 seconds.
  • The projectile does not persist once it hits an enemy. It is only able to affect one enemy for each use.
  • The projectile is blocked by barriers, Defense Matrix, and Particle Barriers. It can be deflected by Genji’s Deflect ability.
  • 10 second cooldown. The cooldown begins after the ability is cast.
  • There is a .25 second wind-down time after Pacify is launched.
  • During the wind-up and wind-down times, Mercy is not able to use her staff, her pistol, or Resurrect. She may use Guardian Angel, however.
  • Does not affect non-player entities and constructs (turrets, Rip-Tire, Supercharger, etc).

This ability came with some concept art:

This ability functions similarly to Ana’s Sleep Dart in its cast, but the projectiles and effects are very different. Because Mercy is supposed to have a low skill floor, the hitscan weapon type and large hitbox make the ability mechanically forgiving. As a result, Pacify’s 33% damage decrease for 4 seconds is less effective than Sleep Dart’s disable for up to 5.5 seconds.

That said, Pacify still has a lot of playmaking potential when mastered. For example, when used on a target that has been Nano-Boosted or Supercharged, the damage increase that Nano-Boost/Supercharger has on that target is completely negated for the next 4 seconds (100 DPS -%33 = 67 DPS + 50% = 100 DPS).

Pacify also drops Widowmaker, Hanzo, and Pharah’s weapon damage below the 2-shot (body shot) range for 200 HP heroes. If used at the right time, it can potentially save an ally’s life.

Caduceus Staff:

  • Healing beam’s rate of healing remains at 60 health/second.

Angelic Descent:

  • For the first one second of Angelic Descent’s use while airborne, Mercy does not lose altitude. After one second of this, Angelic Descent decays into the featherfall currently provided by Angelic Descent.
  • This effect can only be used once while in the air. If it has been expended, Mercy must first make contact with the ground before using it again.

Mercy’s midfight mobility potential will not be increased by much through this change, but it does allow Mercy to reach more positions prior to engagements without help from allies. Moving from ledge to ledge is made easier by these changes.

Credit goes to Brawl for the Resurrect and Pacify ability ideas and to Silawatsi for the modifications for Angelic Descent.

You can find the full document of ideas here:
The Drawing Board - Google Docs

And you can find the survey results here:
Drawing Board Survey

So that was a post. I wonder how this thread will go compared to the last few I made? Let’s find out!

August 28th, 2018 edit: This thread is now the most popular thread of all time.

October 19th, 2018 edit: Slyther0829 put a lot of effort into creating and analyzing a survey regarding people’s opinions on Mercy. It is the most comprehensive and detailed survey I have ever seen on the subject, and I strongly suggest checking it out and showing some love:


February 6th, 2019 edit: Made some changes to make the post more relevant to today and also tidied up the edits a bit.

June 6th, 2020 edit: 2000 Likes!


This is unacceptable… The image only have 5 teammates, not 6. the team must have a stealth Sombra scouting the area.

But as always nice quality thread! :sweat_smile:


Here we go! Another nice and big Titanium thread!



Well thanks for the book about how mercy still. Needs a rework


I feel like it should be mentioned that I started doing a more comprehensive list of who “we” is. There’s just too many names for one person to do in a timely manner, but what I found is that the vast majority (of 15 that I saw, only 1 wasn’t trawling for infractions) of moderators who have read counts on par with forum monkey’s like me are the ones that lock threads or banish them to the Mercy thread. The ones looking for violations of the CoC are the only ones who technically read feedback, but that’s more to catch people breaking rules.


Thank you for this incredibly well written post, as always I am astounded at the effort and intelligence you pour into your writing.

After voting on the suggested abilties/ult, I have to say that I am quite in love with the Solution you have put forward. A balanced mass rez that promotes tempo rezzing and has ample counterplay, and a new E ability which requires skillful usage and doesn’t outshine other support heroes.

Also, the removal of Valkyrie gives Mercy weaknesses again, and moving rez off on an E ability further enforces her to being more niche and not a must pick.

Great work <3


Why are there 4 dps though?


Tracer is in the back fighting Reaper.


That is my point. The people reading the feedback (“we”) are people who have no authority in balance changes, or might not even be in contact with the developers at all.


It’s Quickplay.

Edit: And it’s a good Quickplay match too; most of the team is on the payload for once.


Ok. *picks Torbjörn * tORbjORN


Omg I just finished resurrect section…


Would you remove damage boost to compensate for Mercy’s higher mobility, new utility, consistent heals, and mass rez returning?


Pacify is such a cool ability idea. Blizz better not steal it for another character.




Too late, Alexstrasza has it in hots


Then it’s too strong. You’re increasing the mobility and utility advantage she has over moira and Ana, and keeping her healing just as strong.


This is quite the post.
Glad the megavoid is practically dead… Hopefully.


I meant for this game. I thought that was obvious.

1 Like