Let's talk about Champ Select

Posted on at 4:55 PM by Moobeat
As a follow up to the post earlier today, several Rioters have taken to the forums to further discuss player behavior in champion select.
Continue reading for a lengthy discussion about the current state of champion select, some commonly discussed solutions, and more.

Here is Lyte's initial post, covering the problems and possible champion select for champion select.
"We want to take some time today to talk about Champion Select. 
1) What are the problems? 
2) What are some potential solutions? 
The player behavior team has been running research on Champ Select and we agree that Champ Select is currently not a great environment and does not set teams up for success. We’ve all experienced Champ Selects that have erupted in arguments and had that sinking feeling that the game is lost before it even started. In saying this, there are plenty of Champ Select lobbies that are awesome and being positive and cooperative in every lobby does help; however, being positive by itself will not solve the problems in Champ Select and we don’t expect it to. 
Solving player behavior problems in League of Legends requires collaboration between us and players and we haven’t done our part in Champ Select yet. As you can imagine, the problems (and any potential solutions) are complicated. 
Something we’re seeing in our research is the influence of context.

Consider a player that’s having a bad day at work and nothing seems to be going right. This player goes home, and loads up some League of Legends thinking, “I really just want to be a Mage today—maybe Annie.” They enter a lobby, and second pick calls “Annie.” The first pick says, “fk, I’m mid, I want Twisted Fate. I’m first pick.” The lobby crumbles in front of the player and he doesn’t even want to throw his suggestion into the mix—it’s a lost battle. The team enters the game, and the player plays poorly… and his teammates yell at him. “You suck at Support, why didn’t you choose a Tank.” The player snaps. He rages back.
But you know what? 
This player's behavior isn’t toxic. He’s just like any one of us--we all have our bad days. 
Let’s break down this scenario into some of the problems that we’d like to solve in Champ Select: 
1) Real-Life Context | This scenario really illustrates how context outside the game can influence behavior inside the game. Traditionally, game studios don’t design or solve for context. Or can they? 
2) In-Game Context | This scenario illustrates the conflict between Pick Order and Call Order. When there are literally no guidelines, at best, half of the players believe in Pick Order and half believe in Call Order—we’ve created a situation where conflicts are expected rather than rare.

3) Time Pressure | From psychology, we know that time pressure sometimes twists context in hostile ways. Players in Champ Select are effectively trying to negotiate with each other over individual goals (i.e, what role I want to play this game) that overlap with team goals (i.e, given this set of teammates, what’s the best strategy for us to win?). Studies suggest that throwing time pressure in there is like adding fuel to the fire—the end result is more disagreements and lower quality of negotiations. 
4) Cognitive Biases | Hopefully davin will talk more about this, but people show cognitive bias in many ways. For example, many of us enter Champ Select thinking we are the best at whatever role we want to play—this is statistically impossible; however, there’s no reason to trust any of the strangers in the lobby. This really isn’t the players’ fault, it’s simply being human. 
These are some major problems with Champ Select that we’ve identified in our research. So what’s next? A lot of players have suggested the following: 
1) Vote Kick | Players want the ability to vote kick toxic players from Champ Select.
2) WoW Dungeon Finder | Players want the ability to queue up for a particular role like “Healer” and “DPS” and placed into a Champ Select with a team 
3) Prisoner’s Island | Players want matchmaking to pair toxic players with toxic players, and positive players with positive players. 
What are some pros and cons to these ideas? Would they work for League?"

A few Rioters then moved on to discuss a few of the points in more details.

Lyte started off by talking about Vote Kick, saying:
"Vote kick is an interesting idea that has been suggested by players for quite awhile; however, what are some of the goals we'd like to accomplish for Champ Select? 
One, we want to encourage cooperation in Champ Select. Two, we want players to have an ability to opt-out if they are 'stuck' with players they perceive are toxic or extremely negative. Vote-kick systems tend to give players an opt-out mechanism; however, they do not encourage cooperation.

In fact, in some scenarios vote-kick systems encourage premade groups to bully the strangers in the lobby into specific roles or champions. Given a scenario with 5 strangers, if 4 strangers happen to agree on roles and the 5th doesn't, the 4 strangers are highly likely to collaborate to kick the 5th. Given a vote-kick system, we are likely to see more disagreements than before, and greatly increase the time it takes to get into a game."
davin, a senior user researcher, shared his thoughts on the "WoW Dungeon Finder" approach, saying:
"The tricky part here is that unlike WoW's Dungeon Finder, we don't have a single way of playing the game. Dungeons in WoW are designed to pretty much fit the tank/dps/heal model.

So when you match people together, you'd need some way of pairing together players who have agreed on a particular strategy or want to play in a certain way. Otherwise you might end up with two people who queued as best-at-Mid, and unless they're fine playing Double Mid, you'd get some pretty similar behaviors going on."
He continued on to say:
"Let's talk about these two points a little. From a queue time standpoint, a standard-meta queue actually wouldn't be too lengthy. When you take into account the multiple ways of playing support (Tank/Support/Kill Lane) and the players who are fine to fill all roles, you actually have around 15 to 20% of players being down for Jungle and Support (depending region and definition of those roles). So queue time may actually be less of an issue than it'd seem based on the popular perception of those roles as less-favored. 
There may be other reasons why players aren't volunteering support in Ranked even though we know they are probably down for it. There's the potential for "have to carry your team!" perceptions pushing people out of the role, or the feeling of not being able to strongly influence what's happening, etc. As a support main, I'm not in agreement with those ideas, but I definitely recognize they're out there :)

Regardless, the dungeon-finder approach relies upon the idea of forcing a single ranked-style meta. And even though a ton of players do play the bruiser-top, mage-mid, carry/support bottom, jungle layout, there's successful strategies that rely on assassins mid, jungle-mages, carry/support top, etc. I think a decent amount of players wouldn't be down with a queue that only allows for one team layout."
Sort of on the same page, Lyte shared his thoughts on role preference indicators.
"Role preference indicators are an interesting idea. Awhile back, kitae asked me if it we could try a few experiments where players enter Champ Select with some icons that showed their role priority (1st = Top, 2nd = Mid, etc) and see if that improved communication. The idea is compelling because it's a way for players to show their expectations heading into the lobby.

It certainly streamlines some of the communication currently required in a time pressured lobby."

Lyte also elaborated on the problem with Prisoner's Island, explaining:
"When discussing Prisoner's Island, this is an extremely important point. Let's consider a thought exercise: 
1) Let's put 1 toxic player with 9 positive players
- In this scenario, does the toxic player improve his behavior over time, or do we simply ruin the experiences of 9 positive players? 
2) Let's put 10 toxic players together
- In this scenario, do any of the toxic players ever improve their behavior?

A core philosophy on the player behavior team is to make features that help toxic players reform. In many ways, a Prisoner's Island feature encourages the opposite of reform.

Prisoner's Island also creates some pretty weird scenarios for players. When players browse through the Tribunal Ban Inquiries forum, there are numerous players who use excessive verbal abuse and racial slurs in their matches; however, they lack the self-awareness or necessary feedback to understand how negative their behavior is. If these same players are on Prisoner's Island, how many of them would understand why they were there, or how to get out? 
If players don't believe they deserve to be on Prisoner's Island and every other player there is just a jerkwad, doesn't this encourage them to make new accounts to start over, off Prisoner's Island?

If this scenario happens, what was the point of creating Prisoner's Island?"

The Rioters then responded to a few player concerns and questions.

Lyte responded to someone asking about dodge penalties in relation to the possibility of a vote kick system being implemented, explaining why queue penalties are around.
    "We've tested many different queue penalties in the history of League of Legends. If you don't have any penalties at all, queue dodge rates skyrocket. Players leave for many diverse reasons:- perhaps the team composition wasn't optimal
  • - perhaps the opposing team composition was too strong
  • - perhaps they didn't get their role
  • - perhaps they didn't get their champion
  • - perhaps someone was being toxic
  • - perhaps someone was AFK
  • - perhaps the Captain didn't do the drafting the way the team wanted
  • - perhaps a perceived OP champion was chosen by the enemy team
  • - perhaps a perceived UP champion was chosen by a teammate

    What happens is you have more lobbies that end in a queue dodge than actually enter a game. However, when we introduced stronger penalties, queue dodge rates dropped by 50%. Changing and tweaking queue dodge penalties is an extremely tricky science. On one hand, we want players to be able to have an opt-out mechanism to get out of games that are definitely not going to be fun; however, if we have no penalties at all, the actual time it takes a player to hit "Play" and enter the game after several lobbies increases dramatically past 10 minutes. Micro-managing Champ Select lobbies really isn't that fun, and we don't want to encourage it."
When a summoner suggested pick order should be based on Honor score, davin responded:
"Basing pick order on Honor as you suggested is an interesting idea, but ultimately would likely result in similar ambiguity. Part of what we're interested in is removing system-level affordances that create ambiguity and allow for bad behavior to manifest. Which is very much about encouraging positive behavior, not creating new punishments."

Another summoner shared his thoughts that the problems with champion select stem from the player base, not the actual selection system itself. Status Kwoh responded to this by saying:
"I actually disagree. I think that the fear and concern that players have around role and champ select stems from a very natural place of wanting to do their best and have the team do their best. 
Let's look at the typical situation. I drop into champ select with the firm knowledge of being awesome at my favorite role or champion but with zero knowledge of how good anyone else is at the other roles. If I have a known, high, level of confidence in my ability but zero knowledge of another player's ability, how much confidence and trust do I have in them taking it over me? 
That thought process is usually where this behavior stems from. This is why we need to find strong and healthy ways for the game to smooth these situations out.

How can we help people feel more confident and build trust with their teammates pre-game so that players don't feel the pressure that can lead to negative outbursts? This is the sort of discussions Lyte and I have been having a lot here at the office."

Lyte capped off the discussion by saying:
"There is a probably a misunderstanding here! 
We want to try stuff. We're going to try some pretty risky and crazy experiments to try to tackle problems in Champ Select. But, we wanted to have a discussion with players about the most commonly discussed ideas we've heard since the player behavior team formed. 
We've never had a serious discussion talking to players about the benefits (and costs!) of the most popular community ideas such as vote-kick or Prisoner's Island, and I've been thoroughly enjoying the dialogue and ideas bouncing around.

We've been taking a lot of notes about what players care about, what they don't care about, and what types of ideas are worth more brainstorming. This is a problem we'd like to try to tackle on the player behavior team. We just don't know if the experiments are a combination of multiple ideas, a modification of vote-kick, or a completely new concept like Positive Island + something else instead of Prisoner's Island, but we are most likely going to try an experiment or two."
Status Kwoh also ended by noting:
"We agree that focusing on positivity and less on punishment is what we should be doing. We've already begun making that shift focusing on features that work more specifically on reform and paths to sportsmanship/positivity for the playerbase."

No comments

Post a Comment