Greetings! I like analyzing things, so that’s what I’m about to do with a few numbers regarding the six support heroes in Overwatch. In case the word, “numbers” didn’t give it away already, I’m going to be talking about the balance of the aforementioned heroes… and maybe a bit of other stuff.
Let’s start vague. Over the past week, here are the Competitive pickrates of the six healers according to Overbuff:
Overall Support Pickrates
Hero Pickrate Ana 9.80% Moira 7.29% Mercy 5.18% Lucio 4.68% Zenyatta 4.20% Brigitte 3.52%
Given this data, there doesn’t appear to be anything wrong in regards to the balance of supports. While there is a large difference between the most-picked support and the least-picked support, none of these heroes appear to have oppressively high pickrates or pathetically low pickrates. From this set of data, we could conclude that the six supports are fairly well-balanced. We might also conclude that off-healers generally are less powerful than primary healers.
Let’s get less vague. Here are the pickrates of those same heroes throughout each rank:
Support Pickrates in Each Rank
Hero Bronze Silver Gold Platinum Diamond Master GM Ana 5.11% 6.73% 8.48% 10.54% 12.00% 14.06% 13.00% Moira 9.84% 9.96% 8.82% 7.16% 5.18% 2.16% 0.86% Mercy 7.99% 7.51% 6.43% 4.74% 3.26% 2.42% 1.41% Lucio 5.45% 4.19% 4.27% 4.09% 4.79% 6.43% 9.94% Zenyatta 3.18% 3.31% 3.55% 4.26% 5.08% 5.43% 7.42% Brigitte 3.48% 3.51% 3.47% 3.33% 3.44% 4.06% 5.62%
There are several things I want to highlight here. First, this data set shows us that the off-healers are well-balanced amongst each other throughout all ranks. In no rank does one of them have a pickrate greater than or equal to double either of the other two off-healers. It also shows us that the off-healers are pretty well-balanced overall when looking at their high-tier pickrates.
This set of data disproves the previous set of potential conclusions. If we isolate this set down to the very high tiers, where viability starts taking substantial precedence over fun in hero choices, we see that the three off-healers are more powerful than two of the three main healers by a long shot… which brings me to the main healers. Their pickrates really start to seem problematic at Diamond, and they only get worse from there. In Diamond, Ana has more than double Moira’s pickrate, and nearly quadruple Mercy’s pickrate. Moving up the ladder, this trend worsens until we reach GM, where Ana has more than 9 times Mercy’s pickrate and more than 15 times Moira’s pickrate.
It doesn’t take a mathematician to realize that there’s a problem here. Either Ana is insanely overpowered, or Mercy and Moira are terribly underpowered. A hero doesn’t hold a pickrate greater than 9 times the pickrate of their counterparts in GM unless either they are too powerful, or those counterparts are too weak.
Now, I’ve seen some people on this forum argue for the current state of supports to remain as it is, and usually the argument behind that notion is related to skill. The argument tends to be that “Because Ana is a high-skill hero and her counterparts are not, it is okay for her to be more powerful than them in high-tier play. If Mercy and Moira were more powerful, nobody would pick Ana because they offer more reward for less skill”.
This notion has merit to it. While the way it is being used is heavily flawed, the concern that a powerful hero with low skill requirements will overtake a powerful hero of greater skill requirements isn’t unrealistic. The problem is that the concern isn’t being used realistically in Overwatch’s current state.
I could argue against complacency with the current state of supports all day using theoretical arguments, but there is a better way to prove this point: Statistically.
First, let’s refine some terms into more workable parts. Generally, the phrase “high-skill” tends to be synonymous with “difficult to play”. Literally, no hero is difficult to play; anyone who can use a computer can select a given hero in the hero select screen and stay on that hero for the match, but I’m fairly certain that “difficult to play” is not meant to be said or taken literally. I want it to be taken literally for our purposes, so I think that a better terminology would be “difficult to climb while playing as”. If a hero requires more skill, then it will be more difficult to climb with them when playing as them. Likewise, we can refine _low-skill" down to “easy to climb while playing as”.
Before I continue, a quick disclaimer: In saying what follows, I am not in any way trying to insult, shame, or criticize players based upon hero main, rank, or any other personal skill/preference measurements.
First, let’s grab some stats from Jeff Kaplan himself:
The image showing the stats is missing because I can’t post images/links, even if they are quoted from someone else. You can click on the quote to see the full post if you like, but for your convenience, I’ll throw the stats in the table below:
Rank Distribution
Rank Playerbase Proportion Bronze 8% Silver 21% Gold 32% Platinum 25% Diamond 10% Master 3% GM 1%
While this information is nearly 10 months old, it is worth mentioning that the distribution barely changed between March 3rd, 2017 (the only other time the distribution was posted by Blizzard), and February 20th, 2018 (the quote you see above). I think it is safe to assume that the distribution of players throughout the ranks today is approximately the same as the distribution given to us on February 20th.
What am I going to do with these numbers, you might be wondering? I’m going to do an unnecessary amount of bit of multiplication. I’m going to multiply the the proportion of Overwatch players in each rank by the pickrate of each of the main healers in that same rank. This will give me the proportion of Overwatch players who are in each rank and who play each hero.
For example: Ana’s pickrate in Bronze = 5.11%. The proportion of Overwatch playerbase in Bronze = 8%. The proportion of the Overwatch players who play Ana and are in Bronze = 5.11% * 8% = 0.41%. I will repeat this process for all ranks and all three main healers, and fill in the values in the table below.
Playerbase Proportions (Main Healers Only)
Hero Bronze Silver Gold Platinum Diamond Master GM Ana 0.41% 1.41%% 2.71% 2.64% 1.20% 0.42% 0.13% Moira 0.79% 2.09% 2.82% 1.79% 0.52% 0.06% 0.01% Mercy 0.64% 1.58% 2.06% 1.12% 0.33% 0.07% 0.01%
Side note: This model underestimated Ana’s overall pickrate (the sum of all proportions should be the same as the overall pickrate) and overestimated Mercy and Moira’s overall pickrates. This leads me to believe that a larger proportion of the Overwatch playerbase is above Gold today than what is shown in the table provided by Jeff Kaplan in February. Just an interesting tidbit.
Anyway, I now have a yet another table of values. What can I do with it?
I can calculate the likelihood that a (insert main healer here) player will be at/above/below (insert rank here) with these numbers. For example, there is a 9.78% chance that a randomly selected Moira player will be in Bronze.
Let’s use this to calculate the probability of some other things…
The probability that a randomly selected Ana player will be in GM is 1.46%.
The probability that a randomly selected Moira player will be in GM is 0.12%.
The probability that a randomly selected Mercy player will be in GM is 0.17%.
Statistically, you are 12.2 times as likely to be in GM as an Ana player than you are as a Moira player.
Statistically, you are 8.6 times as likely to be in GM as an Ana player than you are as a Mercy player.
Let’s do another set of calculations.
The probability that a randomly selected Ana player will be in Silver or Bronze is 20.40%.
The probability that a randomly selected Moira player will be in Silver or Bronze is 35.64%
The probability that a randomly selected Mercy player will be in Silver or Bronze is 38.21%.
Referencing back to one of my earlier paragraphs, a “high-skill” hero can be defined as one who is difficult to climb with while playing as them.
So. If Ana is truly the “high-skill” hero people on these forums are making her out to be, if it is supposedly harder to achieve a certain benefit, and by extension, rank, when playing as Ana than it is when playing as Mercy or Moira, why do the statistics indicate that it is significantly easier to reach GM as Ana than it is with Mercy or Moira? Why do the statistics indicate that a player can cut their chances of staying in Bronze or Silver nearly in half simply by switching off Mercy/Moira and playing Ana? Why are players who pick Ana statistically more likely to be above the median rank than players who play Mercy or Moira?
Ana isn’t the monument of skill some people like to make her out to be. For the amount of effort she requires, she offers significantly more reward than the other two main healers. For a Mercy or Moira player to reach the same impact as an Ana player, they need to be significantly better than the Ana player.
“Skill” isn’t an integer. It’s a ratio. It’s the amount of effort put in against the amount of reward given for that effort, in both the match-to-match basis and the big picture throughout the ladder.
If I put 100 hours into learning Widowmaker on one account and climb from Gold to GM, and then I get on another account and have to spend 200 hours learning Winston make the same climb, Winston, is in fact, the more difficult hero. I needed to put more effort in as Winston to reach the same reward. Thus, of the two, Winston would be the high-skill hero.
If I had 100 hours to spend on learning Ana, Moira, or Mercy, with my only goal being to climb as high as I can in Competitive play, I would put that time into learning Ana; she provides the highest chance of reaching each and every rank above the median. If I spent those 100 hours on Mercy or Moira, I would have a lower chance of reaching any rank above the median, and I would need to spend more time perfecting my play as either of them to reach the same rank as I would have from playing Ana for the first 100 hours.
The argument that a powerful hero with low skill requirements will overtake a powerful hero of greater skill requirements applies today to Mercy, Ana, and Moira, but not the may many think. I see it most often applied backwards, because players seem think that the skill of a hero never changes (unless your name is Doomfist, but we’re not going to to talk about that here). In reality, because skill is a ratio that depends upon the effort and reward of a hero, power holds great weight in how skilled a hero actually is. The fear of Ana, a “high-skill” hero, being overtaken by Mercy and Moira, “low-skill” heroes, is blinding people from what is actually happening. Ana, the low-skill main healer, as shown by the statistics, has overtaken Mercy and Moira, the high-skill main healers, as shown by the statistics.
That is all.
