The final week of regular season play has ended, and the teams know where they stand heading into the playoffs.
Although many teams have experimented and been “playing for fun” the last month or so, the playoffs will finally see the teams playing for real when prize money finally comes into effect. Although it isn’t as large as many of the former professionals are used to, the opportunity to compete for a reward will most likely bring out their best effort.
To sum up the season, NexusDefender will share a statistical breakdown of the players at three main positions: tank, healer and assassin.
As a note, the offlane and flex roles aren’t initially reported on because of the difficulty to statistically measure their impact on a match-by-match basis.
I began tracking these statistics around week 7 of HeroesLounge, so I will say this: they aren’t 100 percent definitive. The amount of time it takes to import all the match data from HeroesLounge to my spreadsheets and run it is pretty long, so I can’t track everything, and there may be matches I missed.
For fairness and simplicity, I tried to only include players that I had 10 games recorded of them playing the main healer role on the team. I did include Legacy, although I only had a few games for him, so his stats may be inaccurate.
I created these tables to show a players effective healing and takedown/life (I prefer lives (deaths+1) because you can’t divide by zero) rate. For more information about how I got these metrics, read below.
According to this data, Hooberman had the most successful season for a support player this year.
Hooberman had an effective healing percentage of 0.398, meaning he healed for approximately 40 percent of the damage the enemy team did to his. TigerJK and Darkchimaera both got close to this number, but Hooberman’s Takedown/life ratio also beat everyone else by nearly 1.5.
Darkchimaera is considered by many players to be the best support player in North America, and his rankings are high as well. In addition to being in the top 3 of both categories, his team, Looking For Work, was undefeated throughout the regular season.
KilluZiion and ElhayM both also appear to have had solid performances, earning positive results throughout the year.
Some background from the man behind the cloth…
For context, I want to share a little background about the stats I have and how I found them.
A few years ago, around the time of Heroes 2.0, I was trying to put together a Heroes of the Dorm team that had a fighting chance to compete in large collegiate tournament.
I was naive in this hope, mainly because neither I nor anyone at my school had the experience and skills ready to play at that level. However, I began to wonder if it was possible to increase our win totals through numbers (I was inspired by the movie Moneyball).
Long story short, I tracked statistics from Heroes of the Storm for months. I logged a variety of games starting with my own, then moving on to my teammates. Finally, I logged HGC matches exclusively. About 300+ total matches of base statistics went into my first batch.
When I began looking at the stats and their correlation to win-rate, I was baffled. There were almost no stats in the game that inherently indicated a winning formula. The only one that showed a strong positive correlation was experience gained, which should have been obvious. But numbers like raw damage, healing and kills didn’t seem to matter as much as I would have thought.
After digging deeper and trying to come up with some of my own metrics, I finally found a few that looked very promising. The first I called “effective healing.” This is also the main metric I used for analyzing Division S supports.
effective healing= amounting of healing done by healer/amount of total hero damage by the enemy team
I came up with this formula because it isn’t enough to measure the raw healing by a player. There are too many variables in the game to make that effective. One example is match time. If one winning player is healing less than a losing one, but the winning player is winning in 15 minutes, whereas the losing player is losing in 25, you’ll get misleading numbers.
Another variable is the variety of healers. Some supports, such as Uther, may not share the raw healing potency of a Lucio, but he does other things to help the team win, particularly in team fights.
Effective healing is important, because it shows how effective the heals are. In a fight, is the healer doing enough to keep his team alive until the fight in over? That could be sustaining through massive amounts of damage, or it could be landing a nice stun to win the fight quicker. Simply put, effective healing means how the sustain helps keep your team in a fight longer than your opponents’ healer.
According to my personal data (which now probably sits at around 500 games), this metric will predict the winner or a match 80-90 percent of the time at all levels of play.
Division S playoffs for North America begin July 20.