does riot actully have any numbers they will show the public about toxicity?

JAGMAG·Aug 7, 2015·3 votes·375 views

for real they talk about only 2% of the game population is toxic and that most reform after punishment were do they get those numbers thou i call bull shit

9 Comments

Deep Terror NamiAug 7, 20151 votes

Why would they bother responding to you when you already think they're lying and won't believe what they tell you?

CokcMagicAug 7, 20151 votes

[{quoted}](name=Valkyrie Witch,realm=NA,application-id=cIfEodbz,discussion-id=xhUp8dnX,comment-id=0000,timestamp=2015-08-07T05:52:55.337+0000)

Why would they bother responding to you when you already think they're lying and won't believe what they tell you?

Though I understand what you are saying, I also understand where this is coming from. If I play five games a day, at least 2-3 of those will have had truly toxic or trollish behavior and almost certainly will have had a disconnect or AFK'er on either my team or the opponents. Which is definitely not 2%.

JubbinautAug 7, 20152 votes

No, but it's also not very high. Looking at both cases (and assuming you're not toxic/trollish/prone to afks yourself):

Out of 5 games, you'll (usually) see 45 other players. If you see a troll in two of those games, you're looking at about an average of 4%. If it's three, that skips up to 6%.

Which means that given the team-based nature of the game, a fairly small number of players comparatively can skew our perceptions of the general levels of toxicity. One of the biggest points of confusion, I think, is that Riot tends to talk in terms of toxic players (or players expressing toxicity - I'll come back to this), while players tend to talk in terms of toxic games.

Here's where things get weird, though: most of the toxicity witnessed in games doesn't come from toxic players. Rather, it comes from players who are usually rated "positive" or "neutral," but may be having a bad day, or be on tilt, or what have you.

Lyte gave a talk at MIT a couple years ago, where they showed some interesting data. This was one of those data points (link should skip to about the 5 minute mark, where they discuss this data):

Behavioral Profile Representation (by type of Player) - aka, how many players fall into which category: Toxic - 1% Bad - 7% Neutral - 14% Good - 78%

Toxicity Representation (by type of Player) - aka, how much toxicity comes from which category of players: Toxic - 5% Bad - 8% Neutral - 10% Good - 77%

Now, I don't want to focus on those numbers, but they help illustrate the point - all players exhibit toxicity. If we can assume that even half of the players demonstrating toxicity in your games are usually in the "Good" or "Neutral" categories, that cuts the number of "Toxic" players you encounter to 2-3%.

The strange problem of perceptions this creates is fascinating to me, so I hope you understand that I'm not trying to downplay your experiences, here. Rather, my intention is to give you a little more context to understand how your experiences and Riot's statements can both be correct.

JAGMAGAug 7, 20151 votes

the reason i dont believe them is because they dont give us any preprocessed info/stats they just go and give us a number they dont explain their methods they dont explain were and how they got the number or how they drew their conclusion from them there just like hey guys heirs some numbers we came up with.

for example any research you do in anything similar your required to extensively explain your methods Riots like fuck you its dangerous to give you numbers you might not know what they mean and get confused and numbers could hurt your head and stuff just trust us we have a phd in psychology.

JubbinautAug 7, 20151 votes

{quoted}

the reason i dont believe them is because they dont give us any preprocessed info/stats they just go and give us a number they dont explain their methods they dont explain were and how they got the number or how they drew their conclusion from them there just like hey guys heirs some numbers we came up with.

That's because we're "just" customers/clients.

for example any research you do in anything similar your required to extensively explain your methods Riots like fuck you its dangerous to give you numbers you might not know what they mean and get confused and numbers could hurt your head and stuff just trust us we have a phd in psychology.

They do share their data with plenty of 3rd party organizations. I know there's been a number of schools that have used Riot's PB data for their own studies, formally published in peer-reviewed journals.

I don't have any titles, though, because I've never really been invested enough to find them.

JAGMAGAug 7, 20151 votes

well if we are just customers/clients then why dont they stop with the lie of calling us a "community" and just admit they really dont give a fuck what we think just as long as we keep playing the game

also i must look for these studies x.x

JubbinautAug 7, 20151 votes

{quoted}

well if we are just customers/clients then why dont they stop with the lie of calling us a "community" and just admit they really dont give a fuck what we think just as long as we keep playing the game

My point is that you don't give raw data to a bunch of laypersons. Without the understanding of how to process or analyze that data, it's effectively meaningless.

Instead, you give the results of that analysis - which they do.

also i must look for these studies x.x

Feel free. I think Lyte gave the title of one a few months back on Reddit. You might search his post history for it.

JAGMAGAug 7, 20151 votes

whats to say i cant process the data? as the saying goes figures dont lie but liars figure (this basically means the figures can be slanted to say whatever you want them to say if your cleaver enough)

if you dont believe you can twist figures or get false readings i suggest you look up a i dont think it was a legit study just more of something to prove a point were a guy gets people to sign his petition to ban water by calling water dihydrogen oxide (means h2o) and talking about how this stuff get in our rivers and it soaks into he ground and gets in the plants and needs to be banned and people are like oh that's horrible lets ban it.......so that guy could make a study saying people support the banning of water and to were that's not true....its true..... so id like just to see how they conduct there research and how they get their numbers

JubbinautAug 7, 20151 votes

{quoted}

whats to say i cant process the data? as the saying goes figures dont lie but liars figure (this basically means the figures can be slanted to say whatever you want them to say if your cleaver enough)

Nothing. But you're one out of 27 million daily players. Do you really think dumping that kind of raw data (seriously - more than 27 million unique data sets per day) on a large number of random people is going to do more than confuse most of them?

As a wise man once said "A person is smart. People are dumb, panicky, dangerous animals."

if you dont believe you can twist figures or get false readings i suggest you look up a i dont think it was a legit study just more of something to prove a point were a guy gets people to sign his petition to ban water by calling water dihydrogen oxide (means h2o) and talking about how this stuff get in our rivers and it soaks into he ground and gets in the plants and needs to be banned and people are like oh that's horrible lets ban it.......so that guy could make a study saying people support the banning of water and to were that's not true....its true..... so id like just to see how they conduct there research and how they get their numbers

To be fair, that's an invalid comparison, as he wasn't misrepresenting his findings - he was presenting his raw data in such a way as to be intentionally confusing. Analysis of that data would actually have removed confusion.

I'm quite aware of how figures can be twisted to give false appearances. Most of those methods are fallacies - most commonly, the C = C fallacy (the fallacy of conflating correlation and causation). As best I can tell, Riot actively avoids these.

The only real misrepresentation I could see being possible with what they've given us is false-flagging - that is to say, modifying the criteria of categorization to make things appear differently than they are. But there's no real motive for that, so far as I can tell, and their claims tend to fit player experiences pretty well (once the two are made comparable, anyway. I talked about that in my post to Sir Nic, above).