Game Gabbin' News

Bappie · **Posted:** Wed May 27, 2015 5:30 am

New Rhythm Tengoku (Heaven) game.
https://www.youtube.com/watch?v=46kfIdcTERc

Also the new Wrestler.

Does he remind anyone else of this guy?

AquaBat · **Joined:** Tue Jul 15, 2014 9:26 pm **Posts:** 2373

SaintCrazy wrote:

Posting this here rather than the League thread because it has greater implications for online gaming and online communities.

Jeffrey "Lyte" Lin of Riot Games (my favorite Rioter, because he has a PhD) did this talk about player behavior and the new machine learning system LoL is using to give feedback and even ban out toxic players.If you've got a spare 30 minutes and you're interested in the psychology of how people behave on the internet and online multiplayer games, i highly recommend you check out this video.

http://www.gdcvault.com/play/1022160/Mo ... ing-Player

Also, here's an article about the automated ban system. Really cool to know that people who are extremely offensive in chat can now be smited by Skynet. The future is now

http://arstechnica.com/gaming/2015/05/r ... ds-trolls/

Apparently this system is now banning people for saying something as silly as "gg easy". So much for the future.

Syobon · **Joined:** Mon Jan 24, 2011 10:34 am **Posts:** 15027

Oh yeah, I heard about that. Best part is I saw people on reddit actually defending that. Fucking hilarious.

BurntToShreds · **Posted:** Wed May 27, 2015 1:36 pm

In that video I saw that Riot and Reddit are AOK with shadow banning people. While you could use it on people who are actual dirtbags, Reddit has shown that they are not above shadow banning people who have different opinions. I wonder if Riot will ever use shadow bans in the game chat or their own forums and I wonder how fast it will be abused?

SaintCrazy · **Posted:** Wed May 27, 2015 2:47 pm

People who say stuff like gg easy are usually dirtbags anyway (in my experience), and if they get a ban it's probably because they have a history of being rude in games. It is extremely unlikely that 1 game is gonna get you punished unless you're sending death threats or something.

Plus, the majority of punishments that are handed out are stuff like chat restrictions which are pretty tame. You'd have to really fuck up to get a ban of any decent length.

To be perfectly honest I'd rather have a system that's too strict than too light. Its not that hard to not be a dick in chat. Besides, the negative behaviors might not be that bad to everyone, but as long as people report it, that means that people dont want it in their community. I certainly don't.

I don't know much about the shadow bans on reddit, but it seems like it'd be really difficult to implement in game so i doubt you'd need to worry about it there.

Syobon · **Joined:** Mon Jan 24, 2011 10:34 am **Posts:** 15027

A little trashtalk makes you a dirtbag? I guess every one I've ever played videogames with regularly (including myself) is a dirtbag then. Regardless, automatic ban for something like this is stupid, because it ignores context. "gg easy" is often said as a joke too, sometimes by the losing team. Honestly, emerging victorious after a two hour long intense and painful struggle, who wouldn't express their elation with a heartfelt "EASIEST GAME OF MY LIFE".

To your other point, chat restrictions are a pretty big deal in a teamwork oriented game.

I do understand wanting to try to control "toxic" behaviour, but this takes it way too far. Dota 2 has a pretty decent system in place I feel, a report system and a mature language filter is enough to cull the worst of it.

SaintCrazy · **Posted:** Wed May 27, 2015 4:52 pm

Chat restrictions aren't that bad imo. You get a certain number of messages per game, plus you can still communicate with pings. The idea is players have to choose between wasting their messages for trash talk, or using it for strategy. The vast majority of players hardly use chat at all. The main problem i would see would be at really high skill levels where communication might be more important.

As for trashtalk, it depends a lot on the regional culture and what people are willing to let slide vs what people will misinterpret as being offensive (this is the internet after all, what sounds sarcastic or joking to you might not to someone else). The video had a cool point about regions though; the phrase "your mom" is seen almost entirely as a joke in NA, but is a staple of actual offensive language in Korea.

Ultimately the community decides what is acceptable and what isn't, so if you believe that most people don't have a problem with a little trashtalk, you have nothing to worry about. ultimately its up to the players to decide context and who to report. And luckily the system can recognize when people report for no reason, so their reports weigh less.

they hand-reviewed thousands of cases when the system first launched, and the only false positive i know of is the 10,000th or so case where the guy was actually saying abusive things to himself in third person. It isn't perfect but for a system of this scale, it's quite fascinating.

Syobon · **Joined:** Mon Jan 24, 2011 10:34 am **Posts:** 15027

Uh, if it's banning people for saying "gg easy", then that is a false positive in my eyes. Then again, I've heard people supposedly get banned for not following the meta in League, so maybe getting banned is actually a part of the game?

Also, the article you linked doesn't mention chat restrictions, it just says

Quote:

"These harmful communications will be punished with two-weeks or permanent bans within 15 minutes of game’s end," Riot promises.

Which shouldn't happen for any reason imo. Taking a product that a customer has paid for is a serious enough decision that it should only ever be taken by a human, not an automated system. I understand automated bans for free accounts to deal with botters though.

SaintCrazy · **Posted:** Wed May 27, 2015 5:53 pm

If someone gets banned after "gg easy", there's no way they didn't do something else in combination that would male their behavior ban worthy. I think you're overestimating how heavy handed the system is. One instance of trash talk in one game isn't going to cause someone to be banned.

The "harmful communications" that quote is referring to are things like homophobic slurs, sexism, racism, and stuff like death threats. Not trash talk. The insta-bans are reserved for the worst of cases. The percentage of people in the league community that get actually banned is incredibly low.

the "I'm a paying customer, i shouldn't be banned" argument doesn't sit well with me because the experiences of 9 other players per game are involved. its like if someone was at a restaurant telling other customers to go kill themselves; you kick them out of the restaurant for ruining the experiences of everybody else, no matter how good of a customer they had been before. If the behavior is extreme enough it shouldn't be tolerated.

i believe random samples of cases are still hand reviewed by real people, and of course you can send in a ticket to get people to review your case. Checks and balances are there, though i can understand wanting more. It's just difficult at League's massive scale. In any case it's certainly better than the old Tribunal system where other players could judge cases, i think that actually had a higher false positive rate and took weeks to months to actually see punishments roll out.

SaintCrazy · **Posted:** Wed May 27, 2015 6:12 pm

New post for clarity's sake. Here's some more links, that previous article kinda oversimplified things I think.

Here's the official blurb about the system from the League website.
http://na.leagueoflegends.com/en/news/g ... ds-testing

Here's Lyte talking about the player reform system on Reddit:
http://www.reddit.com/r/leagueoflegends ... ng/crfu61p

And here's the rundown of salient points Lyte made on Reddit from Surrender at 20, so you don't have to parse your way through Reddit comments:

Quote:

In response to concerns over this automatic system, Lyte explained:
"I think there's a perception issue with "automatic" systems. For example, when we think about really high security stuff like credit cards, they use a lot of automatic systems but we don't really "think" about accuracy. In some fields of sciences, a 1% error rate is considered exceptional, even if it's in a higher risk field like medicine.
At Riot, we want to make sure players never have to worry about punishment systems if they are neutral or positive. So, any system that we launch needs to have a 0.1% error rate or lower or we'll turn the system off. In first beta test of the new reform system, we were having Player Support teams review thousands of cases and we found that generally there was 1 error in every 6000 cases. In the case that a player did receive a false penalty, the e-mail has instructions and links to contact Player Support."
He continued, replying to concerns about "false punishments".
"I mentioned this in another response, but in our initial test where we hand-reviewed thousands of cases, we saw an error rate of about 1 in 6000 cases.
Given League's scale and massive playerbase, this is still a large number of players affected, and we have systems and teams in place to fix this as soon as it's detected."

As for that hand review process, Lyte elaborated:
"We already did a round of testing where Player Support teams and Player Behavior teams all banded together and hand-reviewed every single case.
Moving forward, we'll use a technique called "random sampling," where we'll random pull a bunch of cases the system and hand-review those which will give us a representation of how accurate the system is doing without reviewing every single case.
With random sampling, we can purposely target more serious punishments and sample them more often because they are a higher impact if they are mistakes."

When asked if there is a difference between a "false" (bad-faith) report and a "not-punish-worthy (good-faith) report as far as the system is concerned, Lyte noted:
"Yes, the system "validates" every report, and false reports are ignored. If the report is generally correct, but the system just doesn't believe the behavior is worthy of a punishment, the reports will still "stick," and could affect the next potential check of the player."
When asked why they chose a "machine learning" engine for this new system, Lyte explained:
"The reason we like a machine learning engine's approach is because it is very difficult to work around. Because every player's report and honor teaches the system about emerging changes in language, the system can "keep up" with new phrases and trends faster than a human team would be able to.
In the future when players can read random chat logs and vote in the Tribunal again on whether behaviors are negative, neutral or positive, these will also teach the system more about behaviors on a continuous basis."

When asked if the system will note any sort of profanity if you end up reported, Lyte noted:
"The system isn't necessarily out to get rid of profanity. If you say things like, "Fuck, missed that skillshot" you will generally be fine. If you say, "Fuck you, you missed that skillshot" you might end up punished because that's targeting someone."
He continued:
"We've mentioned this before, but for 95% of players in the game, they'll never be exposed to a penalty in their lifetime playing League.
If you are neutral or positive, you should bodaciously have no fear of any system."

Syobon · **Joined:** Mon Jan 24, 2011 10:34 am **Posts:** 15027

Quote:

If someone gets banned after "gg easy", there's no way they didn't do something else in combination that would male their behavior ban worthy.

That would be fine if it's true (not much to go on except Riot's word), but frankly I don't see why "gg easy" has to be a factor in the first place. I feel like I need to point out that the issue isn't just that single phrase here, it seems to indicate that they're targeting anyone who enjoys any sort of colourful communication in their multiplayer game.

To be honest though, that's not nearly a big as deal as banning/punishing people for not following the meta, so this is mostly just me shaking my head and chuckling. It's not like my opinion of Riot could sink any lower.

Quote:

the "I'm a paying customer, i shouldn't be banned" argument doesn't sit well with me because the experiences of 9 other players per game are involved.

Well, that's not what I said, I just said that it shouldn't be automated. Using an automated system to flag a user's behaviour for review would be fine. Putting them in a low priority queue and muting them or equivalent in the meantime would also be acceptable.

The comparison to other automated systems with error rates is quite lacking by the way, for starters the comparison to credit cards is invalid because errors in that system are pretty much unheard of (and just think about how many transactions take place worldwide every second). The comparisons to error rate in medicine is completely devoid of context, namely risk/reward. 99% survival rate for a cancer removal operation that only a few hundred have to undergo? Fantastic! 99% survival rate for a flu shot that millions of people get? National outrage.

SaintCrazy · **Posted:** Wed May 27, 2015 8:06 pm

Obviously if the vast majority of players vote to remove "colorful language" from their community by reporting players, it isn't wanted. If the language in question doesn't offend people, it wouldn't get reported and so the system would assume it is, by and large, fine. And it's important to remember that even if someone enjoys communicating a certain way, if it comes at the expense of the enjoyment of others, it isn't okay. If for some reason saying "haha wow you're such a piece of shit, fuckin loser" to other people is really really important to you, play with friends who won't report you for it. Otherwise, if there's other people in your game, don't be a dick. When i see that kind of shit in my games, even if it's not directed at me, it turns my game experience into a negative one.

There's an unfortunate perception out there where people think games, or the internet in general, have this "culture" where it's okay to be a dick to other people, mostly because nobody's gonna stop you if you do. But... it just isn't necessary to act that way and I think a lot of communities would be better off without it, even if it's become the status quo. And the line between "poking fun" and being a jerk is very, very blurry. I think short, temporary bans and restrictions are a small price to pay to get rid of behaviors that drain a lot of the fun out of games for a lot of people. Maybe the dickish stuff people say in chat don't bother you, but they bother a lot of other people, or else the automated system wouldn't pick them up in the first place.

I also don't know of any occasions where someone has been banned for not following the meta. There's no shortage of people complaining about their bans on Reddit or League forums, thinking they were banned for X or Y reason, but I haven't seen a single one with definitive proof that they were banned just for not following the meta. I have seen plenty where a Rioter comes in chat logs and lo and behold, they were actually just a dick and just "forgot" they said terrible things to other people on a regular basis. It is just Riot's word, but I'd trust that evidence more than a random angry guy posting on Reddit.

I don't expect to change your opinion of Riot, but I've been keeping track of player behavior issues in this game for far too long, and I'm finally seeing real progress in removing at least some of the shitheads from the community for a game I enjoy. An automated system is pretty scary-sounding and you're right, there is potential for things to go wrong. But hand-reviewing every case just isn't realistic and the game desperately needed a better system to take care of toxicity.

Syobon · **Joined:** Mon Jan 24, 2011 10:34 am **Posts:** 15027

Quote:

Obviously if the vast majority of players vote to remove "colorful language" from their community by reporting players, it isn't wanted.

If the majority of the LoL community thinks "gg easy" is worthy of a report, then that tells me a lot about them.

Quote:

I also don't know of any occasions where someone has been banned for not following the meta.

Well, I have, and it's always accompanied with LoL players defending this as a legitimate reason to ban someone, so this seems to be the community consensus too. Here is an example.

The OP even says

Quote:

Its completely OK if meta is forced as mandatory thing, i have no problems with that.

and a response from a riot employee about randoming your hero:

Quote:

Randomly selecting your champion walks a thin line between violating the summoners code and adhering to it, in that the player selecting a champion randomly forgo's the option of selecting what the team might need, and is doing so knowing they may not be able to play the champion effectively. It is reasons similar to these that we have disabled the free champion rotation in ranked play.

Which is really confusing to me since why would you even put in randomming in the first place then?

The silver lining seems to be that not following the meta is a grey area rather than straight up bannable. However that probably boils down to:
If you break the meta and have a bad game you're going to get reported , that happens a couple of times and you get banned.

Which is still almost just as bad.

Quote:

But hand-reviewing every case just isn't realistic and the game desperately needed a better system to take care of toxicity.

It is realistic, they are bodaciously drowning in cheddar.

AquaBat · **Joined:** Tue Jul 15, 2014 9:26 pm **Posts:** 2373

Twitch is banning certain games from being streamed:

flame_warp · **Joined:** Tue Feb 25, 2014 6:27 pm **Posts:** 374

...*sigh*

AWKWARD ZOMBIE

Game Gabbin' News

Who is online