Red Post Collection: Headhunter Caitlyn getting more armor, Voice Chat Discussion, DS Swain "not soon or even soon(tm)", Positive Reinforcement in new Tribunal, and more!

Posted on at 9:06 PM by Moobeat
[ Update: Added in a brief tweet by Morello that Dragon Master Swain, a popular fan concept, is "not soon, or even Soon(tm)" and Lyte discussing positive reinforcement in the new Tribunal! ]

Tonight's red post collection features Riot Whist commenting that Headhunter Caitlyn is receiving a few changes to her model, as well as Lyte discussing the team's stance on integrated voice chat and more!
Continue reading for more information!

Headhunter Caitlyn Getting More Armor

[ The WIP texture accidentally pushed to the PBE on 6/25 ] 
As you may recall from an earlier  PBE update, Headhunter Caitlyn received significant changes to her model but also had missing textures on her helm.

Riot Whist commented these were incomplete textures and a finished version should be here soon.
"Actually this was the incorrect UVs in conjunction with a WIP texture that accidently got moved to PBE. 
The finished texture work will be arriving shortly."
He continued, offering more context on Headhunter Caitlyn's extra armor:

They've tried defending it by saying "BIG BOOBS and MINI SKIRTS DEFINE CAITLYN"
My intent when I made the post you're referring to was not that it is what defines Caitlyn, rather it was pointing out that base Caitlyn also has a boob window and a skirt on. I later posted in that thread to confirm that yes her main reads are hat and gun and I was just responding to mention of the other things. That post was intended to provide context and reasons for those elements appearing in the skin. 
Everybody working on the skin is pretty happy to see Caitlyn a bit more armored up - she feels more badass and powerful. I think the version we have now will make a lot of players who took issue with it happy. 
For a little extra context on it: 
She lost armor in production as we worked on ways to make sure she was still readable as Caitlyn (her legs and arms became more bare for the color contrast to help the gun pop). We're changing her because we do agree with the complaints she is getting, exposing skin to make the gun pop was, in the end, not the only way to make her read correctly. 
And this is why we have PBE, it's not just for balance changes. We want this feed back so we can fix red flags before the skins go on sale. We are extremely open to community opinion and I hope the skin update helps prove that. 
I'm also working to try to better word posts in the future so players can more easily understand my points and not accidentally take it out of context."

Voice Chat Discussion

With the recent changes to Riot's stance on third party apps, Lyte has taken the time to post Riot's current thoughts on an integrated voice chat and statistics they have gathered using in-game surveys.
"At Riot, we're pretty hardcore gamers. Many of us have lived and breathed games like Counterstrike for many years of our lives. Given the features games like these have, every so often, there's a heated discussion within Riot about voice chat - should we build it in League of Legends? Would it improve the player experience for the average player? Why do some games choose not to integrate voice chat, while others do? Rioters often debate for hours about the pros and cons of voice chat, so we decided to more deeply investigate the impact of voice chat. Every time the discussion comes up, we add a little bit more research to the mix and learn a little bit more about voice chat's impact on online games. 
Due to the recent rise of 3rd party voice chat applications, we decided to take a moment to share some of our findings and explain why we haven't implemented voice chat in League of Legends, but also why we haven't closed the doors on it being possible in the future.  
Do players want this feature?  
One of the critical things I want to highlight is the difference between integrated voice chat with strangers, versus with friends. We often survey League of Legends players on whether they prefer to use voice chat with friends, and over 79% of players agree that it's a more enjoyable way to play League of Legends. But when you ask League of Legends players whether they prefer to use voice chat with strangers, the agree rate drops to 50%:  1 in 2 players do not want to voice chat with strangers, or don't care. More importantly, 28% of players disagree. When looking at numbers like these, we have to carefully consider that adding voice chat would actually create an unwanted experience for just over 1 in 4 players. Meanwhile, players who want voice chat already enjoy it with friends using other voice options. 
These numbers might seem different than results from online polls or general threads you've seen here on the forums, but there are two important variables at play. One, players that tend to visit online forums or communities are a unique subset of the League of Legends population. Two, online polls or questions about voice chat rarely distinguish the difference between voice chatting with friends versus strangers. If you just ask the question "Do you want voice chat in League of Legends?", most players answer the question through their perspective. Players that prefer to voice chat with strangers say "Yes!", but so do players that prefer voice chat with friends, leading to bloated numbers and inaccurate perceptions. 
What's the player experience like? 
With the recent surge of 3rd party voice chat applications, we were able to do some neat data analysis on the impact of voice chat to the League of Legends experience. We were especially curious as to whether or not these applications created a more enjoyable experience in games where there was voice chat between strangers. 
One of the things about voice chat is that more often than not, you end up with games where only a few of the players opt into the application, and the rest are still using text. In our analysis, we found that text-based communication in games where only some of the teammates were in voice chat had up to 126% more racism, sexism, homophobia, and verbal abuse that we all agree have no place in League. Not surprisingly, we also saw that players who were using voice chat with strangers received 47% more reports than your baseline League player.
There have also been a few concerning studies about gender cues and voice chat in other popular, competitive online games. For example, a published paper from Ohio University showed that just giving a cue or sign that you're a woman on voice chat leads to 300% more negative comments compared to a male voice, or no voice at all
At the end of the day, voice chat with friends is a great experience, and what League of Legends players actually want. Many players already use voice chat like Skype and Ventrilo with their friends and that's awesome. But, a system that automatically or easily puts you into voice chat with strangers leads to 126% more toxicity and 47% more reports even when players can opt-out of the experience or mute each other. 
Hopefully this gives players more insight into how we think about voice chat. 

When asked about the figures of more unsavory verbal abuse, Lyte expanded:
"Like Amazon, Google, Facebook and other tech companies, we've developed in-house predictive models that can analyze chat logs in real-time and assess the level of certain sentiment. For example, the model can detect whether a chat log was generally positive communication, or generally negative communication, while also being able to specifically look at the frequency of certain words that we've labeled or curated as racist or homophobic (like f*g and f*ggot).

Many third party companies also offer similar services, and many game studios use these services to do things like automatic sentiment analysis of social media to determine player perception of a feature, etc. We built one in-house with some experts in machine learning and language analysis because of the nuance required to build such a model for League of Legends, which has it's own unique lingo and slang.

For issues like premade sizes vs all solos, these factors are all taken into account in most systems in League. For example, most research will take into account the current premade size when considering things like the # of reports filed in a game, and whether it's higher or lower than the baseline for that specific premade size vs the general League average."
He continued, explaining the differences in results a close beta voice app might experience vs something like an integrated or default voice chat::
"Results will vary from app to app and implementation to implementation. For example, if a voice chat app is in Closed Beta, the players that apply for beta, install the app and try it with strangers might be more keen on using voice chat with strangers than your typical player. This experience will be completely different than an integrated voice chat solution that enables voice chat very easily for the entire playerbase.

It's pretty easy to determine if you are using voice chat with friends vs strangers because many apps will have markers that are different when you use it with friends vs strangers. For example, if you're using Skype with a stranger, you have to somehow exchange your Skype information over chat first. If you're using Skype with a friend, you already have the friend on your Skype Friend List. If you're in voice chat with someone you've never played a game with before in League, and never play a 2nd game with, more likely that it's a stranger than a friend. A few simple heuristics makes it really easy to differentiate voice chat with friends vs strangers."

Lyte also explained that friends vs not friends in voice chat makes a different in the results:
"The key is whether all 5 people are friends or not. If all 5 players are friends and in voice chat, the experience is better. If all 5 players are strangers, the experience is worse. If some players are in voice, and some are in text, the experience is worse.""
As for how they manged to survey the current LoL playerbase, Lyte explained:
"We did a "live" poll in the game client, so that we had a better representative sample of the playerbase. If we did it on the website, we'd only sample the % of players that tend to visit the forums, which are a completely different demographic than the overall playerbase."
He continued, elaborating on the live polls:
"It's a small pop-up at the end of game that was sent out to a random sampling of players. There's usually no need to send a pop-up to every single player in League; in the social sciences, they've done a lot of work showing that surveying a % of the population can usually get meaningful results that represent the entire population as long as the sampling comes from a representative sample (which is why sampling on the forums would be inaccurate). Many social science studies may survey a few hundred people, but in League we usually survey thousands to tens of thousands and usually do surveys for each server separately." 
He continued, responding to a suggestion of hosting a forum poll to see what players "really" think:
"The forum population isn't an invalid population. It all depends what type of question you are asking, or what discussions you are trying to have. If you're trying to ask the overall playerbase how they feel about voice chat, then it's better to do it in-game to get the most representative sample of players.

However, the forum population is really engaged and passionate about certain topics, and tend to be players who participate in Ranked Modes. So, it's perfectly fitting to engage the forum playerbase on issues like champion balance or issues related to competitive League."
davin also jumped in to comment:
I think instead of the website, perhaps a live poll in the client would be a better idea. The thing I really dislike when statistics are presented in this discussion is that Lyte presented his statistics and then wrote off the possibility that other areas.. such as polls on the forums are invalid without actually pinpointing why his statistics should be taking as valid. In HIS opinion GD visitors are a unique subset of League of Legends and that invalidates statistics taken here. So, I have to ask... If the GD polls are invalid because they are a unique subset, then why do any reds come here for anything? Why are champ reworks developed with the influence of the opinions of a "unique subset of League players?" I'm not saying its the sole source for how Riot gets things done but I mean if our opinion matters and gets used in the consideration of other things, how is a poll on the forums where people clearly spend their non-play time because of their passion for the game not valid?:

It definitely doesn't invalidate statistics taken here, but those statistics are gathered under very different conditions (non-random sample, self-selection, different sample characteristics in that it's folks who visit and post on forums, etc.).

For things that affect the whole community, we can't just rely solely on what a non-representative subset says. It wouldn't be responsible or scientific. But that doesn't mean it's useless or that we should ignore it! We sure as hell will consider it alongside all the other information we gather, though. We're not just slaves to data, we use a data-informed approach (i.e., gather information that helps illuminate a situation).

So when it comes to things like voice chat, research into the situation includes surveys, in-client random sample polling, analytics of what happens in those games, and examination of community sentiment. They all factor in. Multi-modal investigation is the name of the game :)"

As for differences in results across the various regions, Lyte noted
"We didn't see any huge outliers in this research across regions, but there were some nuances in regions like Korea, where many players play in PC Cafes and not at home."

Lyte also commented on the thought of hearing someone's voice  reducing how toxic they would be:
"I've heard it said often that people believe toxicity is linked to how much anonymity there is in a given context; however, several studies and experiments have shown that that's not true. In fact, the key to whether toxicity is present is not whether there is anonymity, but whether there are consequences for actions. 
For example, a lot of news websites and social media sites started adding open comment systems that allowed people to comment on public articles or posts, but only if they logged in to Facebook or other apps that revealed your real name and your profile photo. Very quickly, these sites all realized that reducing the anonymity in these contexts did absolutely nothing in reducing the toxicity of the comments and you still had extremely sexist, racist, prejudice or homophobic comments even when people were forced to reveal their "real" name and profile photo. 
I emphasized "real" name because I suppose it's possible for people to create social media accounts that have fake names for the sole purpose of commenting on public articles and posts, but it wasn't the norm by any means.

One of the greatest misunderstandings is that voice chat will reduce anonymity, which will reduce toxicity."

[ Update ] When asked about different kind of punishments and positive reinforcement, Lyte commented the new Tribunal will feature a way to peer vote on rewards for positive players!
"You're touching on the key point, which is that different players respond to different punishments. We find that on average, about 67% of players improve their communication patterns after just 1 chat restriction experience. If players get chat restrictions and a reform card from the Tribunal that outlines exactly what behaviors led to their punishment, the reform rate goes up to ~75% (varies a bit depending on the server). 
To give a comparison, about 60% of players improve their behaviors after a game ban, and about 70% of players improve their behaviors after a game ban and receiving a reform card. So, on average, chat restrictions tend to do a bit better than game bans when it comes to reform rates; however, game bans tend to force players to create smurf accounts and shift toxicity to lower level players, so there's an additional consequence to using game bans compared to chat restrictions. 
We're also able to do things like restrict players from specific queues, or remove Ranked Rewards, or a combination of any of these penalties. We've also considered things like suspending the ability to use skins or specific champions... but these are untested outside of our research labs so far.

Regarding positive reinforcement, we've always had more plans to add a few more features to Honor. Unfortunately, we pivoted the team to work on Team Builder and want to get back and finish Honor after because we believe it has some of the greatest potential to encourage positive behavior. For example, in the new Tribunal, players that are highly honorable will have a chance to be reviewed in the Tribunal, and if their peers vote them as positive, they will get a reward in-game."

Dragon Master Swain Concept "Not soon or even Soon ( TM )"

With the lack of any details on the next ultimate tier skin, many summoner's have been speculating that the long time fan concept of "Dragon Master Swain" may soon become a reality. 

Unfortunately,  Morello has tweeted that that Dragon Master Swain is "not soon or even Soon ( TM ).
"No news - like I said, it's one we want to do, but it's not soon, or even Soon(tm), more "at somepoint." :)"

Joon Ahn's LoL Work

In a thread discussing Kotaku's recent article about Disney and Lucasarts artist Joon Ahn joining Riot Games,Fafafarani, an Art Producer, commented that Joon has actually been with Riot for a while now!
"He actually has been at Riot for about a year and a half. His summoner name is boogergames, but if you want to play with him you'll need to keep in mind that he pretty much only plays Kat mid 
You've seen his work before, when we had the freljord event
Here is the link to the Freljord image he worked on."
( This and other work by Joon Ahn can also be found on his artstation. )

No comments

Post a Comment