: I haven't played in about a year (Kaisa was last, didn't even know Pyke was coming)
Jango Mage (EUNE)
: Not sure if you were here then or not, but targeted spells now proc minion agro, so that might affect your Panth matches.
I saw that when looking into changes for Panth and Lulu. Annoying but workable.
Rioter Comments
Sharjo (EUW)
: > I'm sure everyone's well aware of the memes that are almost canon at this point. Riven and Annie trip over their skins meanwhile **Gragas**... Has the second highest number of skins any champion currently has at 9. Riven has less than him with 7. There's many champions you could've used here other than him. Like Yorick. Or Urgot. Or Swain. Or Rumble. Or Cassiopeia. The list goes on.
Shit. I'm used to seeing Gragas esq. only. I see a large variety of Riven skins. My bad. I actually would have pegged Swain to have more than Grag because I see a lot of variety for him too.
Rioter Comments
obsesse (NA)
: Although this is completely off-topic, I can't read some seemingly ignorant shit and ignore it. Vaccines don't cause autism. There's zero evidence that they do. The only research that has ever claimed that vaccines cause autism have been retracted due to blatant forgery or other data manipulation. Also, I don't mind at all if we implement a quasi-experimental design. There are very, very few 'true experiments' today in any kind of human studies. Good statistics handle this relatively easily. Accounting for known variance is one such way, and we are particularly good at this. I'm afraid I don't know what you're on about anymore.
That's like saying coming up with a methodology to test whether the Earth is in fact flat is claiming that the Earth is flat. It was just an example to state that while I personally can't find a way to design a manipulative experiment testing the hypothesis that isn't unethical, I have been able to roughly sketch ones that skirt around the ethical dilemmas for a hot button issue. Anova tables have no distinction between type of design. They just do F tests on sources of variances in ratio to error. This is not a calculations problem but a design/logic problem.
: Have you ever been banned?
Nope. Though I have been forum banned 4 times, I have never been banned in game once. Honestly, I wish people would just trash each other, then pull in the other three people, then start fucking with the enemy in /allchat, and then the game was all just one big orgy of shitty trash talking. It'd be more exciting than getting Lee jungle for the 47th Lyte-damn time.
: Can't even believe how childish this game has become (Harassed post game)
So let me get this straight, a guy is annoying you and trashing you all game...and you thought it was prudent to add him so he could talk to you after the game? I'm sorry, but this asking me to feel sorry that you burned your hand after putting your hand in the fire to est how hot it was.
obsesse (NA)
: For sake of argument, statistics is how we determine whether things are related or not. You've basically used the word you're trying to define in the definition. The question is whether or not toxicity and performance are related, and how we know this information. Tbh it sounds like you've never had statistics at all. I don't think anything you said was correct and it's sad that other agreed with you.
Oh no, the question isn't if they are related; I'm willing to take Riot's word that they are. The question is if I purposefully change the rate of toxicity through manipulation (as opposed to letting it vary naturally), will that reliably predict a change in performance or game quality? As far as I'm aware, it would be impossible to determine that without being a completely unethical ass by fucking with your customers' games *for science*, but I'm definitely curious if people can come up with a way to manipulate toxicity directly without being an ass to your customers. After all, I came up with a methodology that ought to test experimentally whether vaccines cause autism in an extremely roundabout way, so I'm sure with enough time and effort, there might be a way to test toxicity influence.
: [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=0008000000010000,timestamp=2017-05-08T19:30:07.728+0000)Why are we tailoring design for feelings? Because this is what UX/UI design and game design IS. I can change your experience by changing a few colors, or the pattern on the ground, or the lighting. If I can change behavior to the positive by presenting certain types of text? You can bet I will. >2) You're confused. If I'm confused, then it's about what your problem is here. They have collected data showing a very strong correlation between toxic behavior and chance of winning. They've never claimed it was the cause, and a strong enough correlation can be actionable -- especially in a case like this where there is no real cost to the players by implementing pre-game text like this. It's in a group social situation, meaning that sometimes -- like we do with addictive drug / lifestyle statistics -- an experiment where you seed toxic players in to test the theory is unethical. Sometimes ethics trumps hard data. So I'm not sure what the big concern is, I suppose.
My concern is two-fold: 1) Adherence to proper methodological principle is paramount. You don't go to the doctor and accept it when he does some shamanistic dance, do you? You don't simply say, "It has been proved. QED" on an analysis test when asked to prove the Mean Value Theorem. You don't get a free pass in saying "murder is evil" if someone asks you to make a case for it. I wish Riot would not cut corners when it comes to player behavior study simply because there may be a crisis in their eyes. 2) I am very annoyed that literally only 1 person presenting an argument against my concern has actually made an argument. Literally everyone else has said, "Well it just makes sense" or heckled me. Those are NOT ARGUMENTS. To hell with the tip itself. I couldn't care less. I'm annoyed that when I'm pointing out, "hey, this is just a correlation. Why should we expect Riot's attempts to mitigate toxicity to accomplish anything meaningful without evidence of influence?" only one person can give an argument but everyone is pretending that they are, and even that one person said something to the effect, "We don't exactly know." I have given a lot of money to Riot and enjoyed playing their game a lot, and I would hate for them to run their game into the ground out of ideological pursuit to crush toxicity by any means necessary.
: and yet here we are, I agree with you but I also am stating that the influx of info can be easily modified and ur cancer and coffee explanation just confuses everyone. 100% of people are human and at the same time 100% of humans die, 100% of the population eats hence if you eat you will die that's what ur example sounds like
Perhaps I should have stuck with Internet Explorer usage vs. Homicide rate. More lulz. But in any case, my point is correlation does not imply causation, but correlation often cited on the boards by both reds and players when talking about behavioral modification. If in fact there is no causal link, then Riot literally just wasting money trying to promote good behavior while also allowing politics and social issues to creep into a game.. And that's the best case scenario of failure. If I am correct that allowing jerkish behavior is a necessary step to developing proper problem solving skills, Riot is literally stunting the ability of their players to deal with toxicity, thereby inflating the rate of toxicity. I would argue this hasn't exactly happened yet because Riot has been focused on banning and reforming, but since Lyte has outlined a planned for proactive behavioral adjustment, this is why I bring up the issue.
: > [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=00080000,timestamp=2017-05-08T18:19:14.098+0000)Still, even if priming does influence a reduction in toxicity, my final question would be "so what?" and then we'e back where we started with whether toxicity in general reduces winrates. Given that they also have millions upon millions of data points showing that the teams with reported players have a lower winrate than games without? I don't think we're back there -- we instead have two compelling data sets that say A: priming in certain ways can influence toxicity, and B: teams with toxic players lose more games. > Sample size is irrelevant; methodology is king. ...in pure statistics, perhaps. In practical statistics? No. I do freelance UX design. I've seen the studies that show that certain colors draw more attention -- and create a higher conversion rate -- than other colors. Am I 100% statistically CERTAIN this is the cause, and not something weird like the thousands of data points all just REALLY liking that one color? No. But the correlation is reliable, repeatable in other tests, predictable in its margin of preference, and effective. I'd be a fool to discount that data as non-useful or non-actionable, since it can reliably increase conversion rates for my clients. Also note that Riot isn't claiming this is the CAUSE. They're instead saying "This is our experiment. These are results which are consistent over hundreds of thousands of games. This has become predictable." It's like science -- even GRAVITY isn't 100% proven. It could be that we don't fully understand it. But we still call it the Law of Gravity, because it stands up to pretty much every test we throw at it. ------------------------------- Practical statistics -- especially where it intersects with the social sciences -- ISN'T EVER GOING TO BE 100%. NO test that relies on human response can be -- humans are tricky to predict, and have millions of emotional and intellectual and sociological variables. We can still derive meaningful statistical results from them. Unless you're saying that the statistics used in psychology, UI design, game design, group think analysis, politics, social planning, etc. are all useless and examples of people misusing your profession.
I'll divide the response into two sections: ideological and the methodological 1) I'm not disputing the correlation. I'm asking "so what?". Are people going to lose more games because of toxicity? Then where's the experimentation to show it causes it? Is it a problem at all even if it's everywhere? Why? How do we know it? Feelings? Why are we tailoring design for feelings? Even if it's a problem and feelings are a suitable reason for a fix, why does riot have to fix it? Why not let the community directly sort it out themselves? 2) You're confused. I don't know how you came under the impression that I require 100% absolute conviction. I require only 95% just like every other statistician. As I've said before, correlation does not imply causation. You found a correlation and it happened to work in your "experiment" of implementing it. Yipee. And? It doesn't even have to be false generally speaking. We evolved a heuristic-style of learning. If correlations weren't the result of causation a lot of the time, we'd be dead a long time ago. That still doesn't change the fact that you need to demonstrate how THIS specific relation works. Let's stick with the gravity example. "Gravity works because this orange is slightly green." That is a completely true statement by the way (p -> q is always true if p is false). If you do not follow the framework of good methodology, your results will be accurate at a different rate than you report. If you hadn't had a hit, you wouldn't be telling me this. Proper statistical methodology is the idea that you would tell me your results regardless of the outcome. Also, since I do moonlight as a psychological researcher, yes, fucking yes. Psychology is fucked. Social psychology is fucked. Did you know the progress of behavioral psychology has not budged an inch for 25 years? The R^2 is still capped out at about 0.4. They still cling to their fads. The first thing they need to do is toss out this idea of multiple intelligences because it doesn't add anything. The polls and studies that predicted a Trump win had Mturk style crowd sourced methodology or something completely foreign. There is a huge myopia in the study of humans, primarily because it's become politicized, but that's another topic.
: 12 weeks cause it has to be verified that all scores are final and all payments are paid and as for the part 2 of your post you are not worth my time I had a good laugh thankyou but I do not want to further communications with you as you come off as a troll and I don't have time for that.
I could say the same to you since you're basically just heckling and not actually attacking my argument. I mean, if you have a paper that blows my entire argument out of the water, I don't see why you wouldn't just win the argument. I've flipped from being extremely conservative in conclusions in the past few years to being more moderate simply by reassessing my own arguments and positions. It's not like I'm shitposting about statistics. My shitposts are low quality Pepes and various memes.
: > "players who abuse their teammates lose more games than those who don't." They posit a statement of correlation, not causation. > and the limit of inference is that you can't conclude influence, much less causation, between two correlated variables. And they aren't trying to. I don't see the problem here.
Like I said, I'm pointing out the tip because it's emblematic of how Riot does not look at toxicity with any nuance and simply implies that toxicity = bad and must be removed by any means necessary. For example: > [{quoted}](name=RoboLions,realm=NA,application-id=3ErqAdtq,discussion-id=UmhRE66y,comment-id=0001000300000000000000000000,timestamp=2015-08-25T22:00:07.136+0000) > > If you are so callous maybe LoL isn't for you. Despite your personal opinion or anecdotal experience, our philosophy, based on interacting with millions of players, is that this behavior is not acceptable to LoL Players. Just because the Wild West was wild didn't mean that the wild things people did in those times were ok, the same is true for the internet and toxicity. Honestly, I don't even think toxic people (who aren't total sociopaths) are even arguing that toxicity is "good." I think it's mostly people who have my view that negative interactions are necessary for growth or that the cure is worse than the disease. I honestly can't look at what Riot has done in the name of player behavior and reformation without seeing Machiavellian social engineering.
: If u use skinned down natural numbers and z scores then you can figure out the outliers for it all and it's how......im not explaining this I have a degree in statcial analysis it's hard to explain it just works so they do have a minor part in it and I know but this kid just spams coffee and cancer
I mean, I can show you my degree in 12 weeks. The university puts everyone's on hold for 12 weeks after graduation (from bachelor's to post-doc). Not sure why, but w/e. Statistical methodology isn't hard. The only thing technical is the differences between distributions and deciding which one is best for the situation at hand. If the set of predictors does not contain a "change in ..." variable via manipulation, you are simply looking at differences and not changes. If you do have a paper that outlines a methodology I'm not familiar with, I would like to read it; gotta keep up with continued self-education and all that.
: > [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=000500000001000000000000,timestamp=2017-05-08T16:32:47.377+0000) > > Oh definitely. I'm not disputing a correlation. > > One possible predictor variable just popped into my head this morning is that toxicity is a manifestation of (poorly) solving disputes under pressure. Simply removing the effect does not change the fact that you can't solve disputes effectively, and thus game quality remains low or lower than it could be. Plus, allowing people to stumble about in poor quality dispute solving and seeing that it doesn't work is kind of how you improve your skills at solving disputes. I'd argue it's necessary. > > Perhaps I'm wrong, but like I said in the OP, I don't see Red posts where they discuss possible nuance in toxicity. Last time I checked, the level of nuance was capped at "most toxicity is from good people, not trolls who want to ruin your day" which would be necessary evidence in my theory since it would be the effect of normal people getting into situations they are not adept at handling and handling them poorly. > > And the reason I'm putting credence into this as opposed to, "well, all those fatherless kids in jail could just be in there because they father beat them on the back of the left cheek every Thursday after drinking only 2-4 bottles of vodka" is because kids who are allowed to beat each other up in a voluntary environment (verbally and physically) tend to turn out better adults AND have better dispute solving skills than ones who are constantly police by adults. > > That's also just a correlation and not an influence (thank God), but it's a correlation that does fly in the face of the correlation Riot presents in terms of conclusion. Right, they have to present it in terms of conclusions...? What else do you draw from analysis? Did you mean causation?
Yeah. Causation. Brain fart. {{sticker:slayer-jinx-wink}}
: > [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=,timestamp=2017-05-08T15:14:57.980+0000)Riot has a loadscreen tip that states "players who abuse their teammates lose more games than those who don't." This pre-game tip system is part of an experiment to see if Priming effects can reduce toxicity. [You can listen to Riot Lyte talk about it here. Starts around the 17:20 mark.](http://gdcvault.com/play/1017940/The-Science-Behind-Shaping-Player) > I mean, what manipulative experiment have you run on this, Riot? Did you systematically fuck with people's games to induce toxicity so you could detect an influence in behavior? Did you hire actors to verbally beat the shit out of teammates to see performance changes? There are no actors hired, nor are they really fucking with games -- this is no different than the priming effect you get from any number of inputs, and is a necessary experimental step before drawing any sort of conclusions -- you DO have to test them, after all. >...the limit of inference is that you can';t conclude influence, much less causation, between two correlated variables. I think you underestimate the sheer amount of data Riot collects. We're talking hundreds of thousands of data points for each variable. Sure, some of them can be thrown out -- but if over hundreds of thousands of games a single consistent variable reliably produces a certain result? There's probably something to that. That is indeed scientifically significant. We're not talking a few hundred coffee drinkers surveyed here. As a statistician, you should understand that the overwhelming correlation there at least merits further study. There's coincidence, certainly, but past a certain point of proving your assertion you have to begin to take a closer look at whether or not there IS some causation effect.
The first point is interesting, and I will review the results if they've published them yet. Still, even if priming does influence a reduction in toxicity, my final question would be "so what?" and then we're back where we started with whether toxicity in general reduces winrates. The second is basically reiterating what I'm saying. The third is just statistical malpractice and a strawman of my OP. Sample size is irrelevant; methodology is king. If the king is shit, the kingdom is rotten no matter how much gold it has.
: So, riot didn't state a causation, and they're supposed to tell people that correlation doesn't mean causation? Solid idea, but you're clearly biased against riot, with the cursing and all. That bias confounds the statement that you're a statistician.
I mean, I can find you some red posts where the underlying premise is just that. If you read my third paragraph, you would have seen that I'm not against the loading tip. I'm just skeptical about the general lack of nuance in red posts when presenting findings about toxicity to the community.
: > [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=0005000000010000,timestamp=2017-05-08T16:07:58.386+0000) > > What? Lol. I'm thinking of a study from the 1950s or so before medicine had been able to show that smoking causes lung cancer. > > But you've completely glossed over the entire point of the example: Smokers drink a lot of coffee for various reasons. Smoking causes lung cancer. The mere fact coffee and lung cancer are correlated does not mean they are causally linked. > > That's all I'm saying in my OP. Ah I am using the chronological view so replies are disjointed. When I switched back, I see that you're just using the coffee/lung cancer is an example of improper conclusions of correlation as causation. However, to the point of your OP, unfortunately we have no idea how their stats are done and what analysis they used. It would be nice to see what data they used and what factors were taken into account. But from personal experience, I believe there is a correlation. In games where my team was losing but banded together rather than clawed at each others eyes over minor mistakes, ended with good team fight turn arounds and winning. When my teammates were at each others throats, they let the other die more often and harass their teammate for dying rather than do everything possible to win the fights.
Oh definitely. I'm not disputing a correlation. One possible predictor variable just popped into my head this morning is that toxicity is a manifestation of (poorly) solving disputes under pressure. Simply removing the effect does not change the fact that you can't solve disputes effectively, and thus game quality remains low or lower than it could be. Plus, allowing people to stumble about in poor quality dispute solving and seeing that it doesn't work is kind of how you improve your skills at solving disputes. I'd argue it's necessary. Perhaps I'm wrong, but like I said in the OP, I don't see Red posts where they discuss possible nuance in toxicity. Last time I checked, the level of nuance was capped at "most toxicity is from good people, not trolls who want to ruin your day" which would be necessary evidence in my theory since it would be the effect of normal people getting into situations they are not adept at handling and handling them poorly. And the reason I'm putting credence into this as opposed to, "well, all those fatherless kids in jail could just be in there because they father beat them on the back of the left cheek every Thursday after drinking only 2-4 bottles of vodka" is because kids who are allowed to beat each other up in a voluntary environment (verbally and physically) tend to turn out better adults AND have better dispute solving skills than ones who are constantly police by adults. That's also just a correlation and not an influence (thank God), but it's a correlation that does fly in the face of the correlation Riot presents in terms of conclusion.
Jo0o (NA)
: I'm all for correlation not implying causation, but in this situation, isn't there a pretty clear cause-effect in place? We've all been in games that have devolved into toxic circle-jerks and resulted in losses, after all. Even outside of intentional feeding or AFKing, we've seen games where people either tilt and get toxic or tilt due to toxicity. Speaking personally, I've observed games lost due to toxicity, losing games won in part thanks to the absence of toxicity (choosing to rally and fight back rather than surrender or stop trying), and I'm sure that I've never witnessed toxicity HELP to win a game. I can't imagine that my experience is unique. It may not satisfy your criteria of an experimental process, but the correlation of toxicity -> poor game performance is infinitely easier to justify than coffee -> cancer. So, you may win a hypothetical thought experiment of refuting Riot's stance that toxicity leads to poor game performance, but ultimately that's a very reasonable assertion on their part and I'm not sure what your counterargument is beyond "I'm a statistician and there may be problems with their evidence". Surely you aren't asserting that toxicity HELPS, or that it doesn't matter.
See my example in another post about my theory of why coffee "causes" lung cancer. The real hair splitting is removal of Toxic players for game improvement. Even if toxicity doesn't actually influence winrate, so long as the variable A that influence toxicity is confined within the player (and THAT is definitely able to be demonstrated just by an observational study), then removing that player will influence the community behavior. HOWEVER, when the system changes to "reformation" and thereby changing players you already have, NOW you must demonstrate a causal link through manipulative experimentation because even if you fix toxicity, variable A is still present, and that will still influence winrate, and thus nothing changes except you've wasted a bunch of money better spent making content or something else.
: Here is how a non-manipulative study like this would work: 1) Pose a question. In this case "Do players that exhibit toxic behavior have a lower win rate than those that do not show toxic behavior" 2) Define "toxic behavior", is this someone who has been permabanned, 14 day banned, received a warning of any sort? Riot doesn't tell us what they use to define "toxic" in this case but they have some threshold. 3) Collect data of players that meet and do not meet the "toxic" threshold. There are enough games played that we can very easily assume normality for the data here. 4) Look at the win rates of players that meet this threshold and compare it to players that do not meet this threshold. This can be done with a simple unpaired t-test. This works if we assume that over a large enough data set, things like afks and inting are balanced on both sides of the toxic/non-toxic line. 5) Now you have a result that can be interpreted. Riot does not say toxic players lose significantly more games so that means you don't need P<.05 to draw their conclusion. You literally just need the average from the "toxic" pool be less than the average from the "non-toxic" pool. Riot has such a large data set at their fingertips they don't need to manipulate anything to draw conclusions. They can also do all sorts of more detailed analysis given the size of this data set, and who knows, maybe they have. They don't make any claim I would consider absurd here, their claim works if the win rate for toxic players is 49.9% and the win rate for non-toxic players is 50%.
And your methodology has done nothing to demonstrate causation or even just influence. It has nothing to do with sample size or type of regression. Without a manipulative experiment, for all we know, the toxic people are all roughly partitioned by variable A. Variable A determines winrate or some variable B that determines winrate or some causal chain of variables that eventually determine winrate. Thus, toxic people would indeed lose more, but their toxicity does not influence winrate. Instead, it's the variable A. Now, it is reasonable from an observational experiment to state that expelling toxic players from a pool or team would increase winrate, but that's because you have implicitly manipulated variable A by removing it. You have to take what you have and make people toxic or non-toxic in order to see if toxicity influences the winrate.
: > [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=00050000,timestamp=2017-05-08T15:37:54.199+0000) > > X: frequency of drinking coffee everyday (in cups per day) > Y: Either dichotomous diagnosis of lung cancer or number of days before failure. > > Do a logistic regression or a survival analysis, and you will find a strong correlation between drinking coffee and lung cancer. The more coffee you drink, generally, the more likely you will have lung cancer. If you're trying to analyze the study I'm thinking you're trying to analyze (the 2010 with the 2016 follow up) You should note that they specifically stated they lacked data regarding never smokers vs smokers and that even the analysis of the 2016 one said: > Although a positive association was observed between coffee intake and lung cancer, it has been reported that high intakes of coffee are frequently associated with cigarette smoking (145), which could contribute to the increased risk of lung cancer. However, we take a conservative responsibility for the results because of the limited study numbers and the absence of analysis between smokers and nonsmokers. (I can't do super script so it's in a parenthesis) Coffee intake and increased incidence of lung cancer could be a weak correlation if they do the analysis between never smokers and smokers with coffee intake and lung cancer. In fact, they also state that: >Our study demonstrates that coffee intake can reduce the risk of oral, pharynx cancer, colon cancer, liver cancer, prostate cancer endometrial cancer and melanoma by 31%, 13%, 54%, 11%, 27% and 11% respectively for the highest versus lowest coffee intake. Furthermore, coffee intake could reduce the risk of liver cancer, prostate cancer and endometrial cancer by 27%, 3% and 12% with an increment of 2 cups of coffee intake. Inversely, coffee intake seems could increase the risk of lung cancer by 118%. However, considering the shortcomings of our research, our conclusions should be carefully considered. Further studies are needed to clarify the potential underlying mechanisms by which coffee intake may reduce cancer risk. Perhaps add that further studies could assess the association among never smokers, particularly for smoking-related lung cancer. IMO, the data on this is just not really the best to make the firm statement that drinking coffee causes an increase in your risk of lung cancer. I am comfortable saying that there appears to be a correlation between increased coffee consumption and incidence of lung cancer. I can't say how strong that correlation is based on the methodology used in the study but I am willing to say there is some correlation. EDIT: This is also a reply half on the below post and your reply to the other that is already quoted above > [{quoted}](name=Z Statistic,realm=NA,application-id=ZGEFLEUQ,discussion-id=QUX6x4pY,comment-id=00020000,timestamp=2017-05-08T15:29:30.860+0000) > > I don&#x27;t have to do manipulative experiments to get the data. All I have to do is look at the lung cancer rates of coffee drinkers vs the lung cancer rates of everyone else. It&#x27;s not complicated. And it seems pretty obvious that it&#x27;s true.. I mean, if you drink a beverage that has a stimulant in its vapors a lot, those vapors are going to irritate the lining of your lungs eventually (much less the scalding temperature of the drink alone), and thus lung cancer develops through forcing cells to repair and replace themselves more often than the organ is used to, and that equals more cases of lung cancer.
What? Lol. I'm thinking of a study from the 1950s or so before medicine had been able to show that smoking causes lung cancer. But you've completely glossed over the entire point of the example: Smokers drink a lot of coffee for various reasons. Smoking causes lung cancer. The mere fact coffee and lung cancer are correlated does not mean they are causally linked. That's all I'm saying in my OP.
: what does that have to do with anything ur trolling straight up trying to get people to argue
You're saying X causes Y because X and Y are correlated, and here's your theory to explain why they are correlated. Let's replace X and Y with "coffee" and "lung cancer". By your logic, having a correlation is sufficient for X to cause Y, and here's a theory to explain why.
: I'm not good with statistics or the linguistics of these types of research. So what is our "coffee" in the current topic?
Either really, though it would probably be easier to conceptualize the relationship with coffee standing in for toxicity and lung cancer rate for winrate. Correlations are reciprocal, so if X is correlated with Y, then Y is correlated with X. "Correlation Doesn't Imply Causation" is said so much that's become a meme or some kind, just like "Welfare is a ponzi scheme." I mean, they're both 100% true, but people just shut it out because it gets said so often by detractors who then don't expound upon the idea like they should.
: Putting aside that the couple studies on this are inconclusive, what is your point?
Correlation does not imply causation. Causation is the limit of what I call "influence" as predicting variables approach 1. In order to determine influence, one must use manipulation, not mere observation. In order to manipulate the predicting variable for toxicity, Riot must purposefully cause toxicity in a manipulative experiment though various ways, possibly the ones I listed in the OP. Coffee and Lung Cancer is just tradition. I could use Ice Cream Consumption and Shark Attacks for a more modern example or Murder Rate and Internet Explorer Usage for a internet style one.
: You sure you're a statistician? You are trying to use an example of two unrelated things as an analogy (which is poor btw, since toxicity is directly related to performance and therefore chance of winning or losing), and SINCE they are unrelated you will most likely find that the % of coffee drinkers is the same in both groups. So, what's your point again?
My point is wrapped up very nicely in your own post. I am using a standard statistical example of two non-causally but correlated variables to show that correlation does not imply causation. I am also simultaneously asking my OP how do you know these things are in fact causally related (or at least influence related). So far, the counter-arguments have been as such: Toxicity causes poor performance? How do you know that? Because Riot has demonstrated that they are highly correlated. How do you know that is evidence they are causally linked? Because toxicity causes poor performance.
: If Riot actually looked at that kind of data, it would be a significantly higher number than the one they fictitiously pulled. It is an impossible number to pull for that matter because there is no control group that "never flames". There's no variable stating "this team has 1" or "this team has 2". In a true study setting that number is impossible to reach. What if both teams are toxic? Was one team more toxic than the other? Or was one team still more skillful despite being more toxic?
You don't need a control group to calculate correlations. Historical control groups will suffice if doing a near -perfect replications If you specify variables too finely, then you have to adjust the critical value for each comparison to be lower. Plus, you end up inflating your explanatory power for the model. It's much better to pick the big guns and just write off what's left over as explainable variance. If I were designing a manipulative experiment, I would first see what is representative of behavior for a team in the population and then divvy it up into three sections: significantly better behaved, mediocre, and significantly worse behaved. then I would label a factor with those three levels and see if manipulating the behavior in the experiment influences winrate. I would then follow up with the same design but expand it into 3x3 where each team is allowed to be any of the three factors and see how that compares while still manipulating. This is of course assuming I am an asshole who is willing to subject people to feeders, being relentlessly verbally abused, and purposefully trying to set them up for a loss (all in ranked and casual of course).
: Um, that's actually pretty easy to find. Do a regression analysis using the data they already have? X: Number of accurate reports of harassment/verbal abuse that have been found on a person based on manual review Y: Win rate of that person They can do further analysis and even go through your win/loss record where you were only reported and compare that with your win/loss record where you weren't reported. If the correlation is very high, then they can say that there's a correlation between abusive behavior with teammates and win rates and the correlation is that the more abusive you are to teammates, the more likely you are to lose the game?
X: frequency of drinking coffee everyday (in cups per day) Y: Either dichotomous diagnosis of lung cancer or number of days before failure. Do a logistic regression or a survival analysis, and you will find a strong correlation between drinking coffee and lung cancer. The more coffee you drink, generally, the more likely you will have lung cancer.
: DUDE you will win more games if you are not toxic its simple the toxic players get removed from the game hence non toxic players can play more and win more than someone who is banned
You will also be more likely to develop lung cancer if you generally drink more coffee.
Lapis (OCE)
: Riot doesn't have to abuse people to get that data, all they have to do is look at the winrates of toxic assholes vs the winrates of everyone else. It's not that complicated. And it seems pretty obvious that it's true. I mean, if you abuse your teammates, they're gonna tilt and not play as well and that equals more lost games.
I don't have to do manipulative experiments to get the data. All I have to do is look at the lung cancer rates of coffee drinkers vs the lung cancer rates of everyone else. It's not complicated. And it seems pretty obvious that it's true.. I mean, if you drink a beverage that has a stimulant in its vapors a lot, those vapors are going to irritate the lining of your lungs eventually (much less the scalding temperature of the drink alone), and thus lung cancer develops through forcing cells to repair and replace themselves more often than the organ is used to, and that equals more cases of lung cancer.
: Why can't it be that they manually researched x number of games where someone was reported, recorded the win rate of those games, and did the same for games where no report was filed?
Survey lung cancer patients whether they drink coffee. Survey non-lung cancer population whether they drink coffee. You will find a positive correlation between lung cancer and coffee drinking, or a positive correlation between not having lung cancer and not drinking coffee. This is lazy calculations and methodology.
Joxcab (NA)
: Can I borrow your tinfoil hat? I need to store some dishes.
Does it count as a conspiracy theory if I'm just saying the people are being lazy and not discerning?
Rioter Comments
: >I'm going to go out on a limb and say you never played with or against her back then. i'm just correcting your wrong assumption and if you wana know about india, go to visit ;) murican freedom, you bomb anything anyway
No, 'murica bombs places that have oil. The ones that bomb anything anyway are currently having a blast in Europe.
: > [{quoted}](name=Z Statistic,realm=NA,application-id=3ErqAdtq,discussion-id=NPvsZWvQ,comment-id=0003000100000000000000000000,timestamp=2017-04-25T18:53:54.021+0000) > > Unfortunately, none of that has to do with the price of salt in India. I wonder how did this comment chain end up talking about salt huh?(here's a hint: You're a problem.)
I'm just a humble Standard Normal Random Variable. I'm not hurting anybody.
: 21 is not divisible by 4 http://image.noelshack.com/fichiers/2017/03/1484660849-jhin-idiota-x2-by-jofegu-daajew4.jpg remove one
Removing 1 wouldn't change 21. The set would simply start at 2.
: fiora was my main. Yes, you would have proced hydra with every hit of her ult (btw, hydra procs would affect NEARBY ENEMIES, NOT THE MAIN TARGET, GO TRY IT YOURSELF, YOU DON'T DO MORE DAMAGE WITH THE AUTO ATTACK PRO, THE NEARBY ENEMIES ARE THE ONES TAKING MORE DAMAGE COMPARED TO ZERO OTHERWISE). And because she's my main, I know her ult would usually (if you caught an adc alone) do only about 40-50% of their hp (around 800-900) damage The reason adc used to die to fiora was: 1) they were out of position (no where near their team) 2) they would panic and not attack stuff to life steal 3) they would, ofc, buy no defensive item (Again, ninja tabi was 10% damage reduction on her auto attacks, and she's an auto attack based champion). 4) they got comboed down by Fiora. This is what Fiora combo was: Q (500 or 550 range btw), into auto attack, into E (resets auto attack) into auto attack, into R. The amazing thing is, if you were among your team mates, fiora would've done even less damage with you with her ult because the damage would spread across your team. Also, if you would've played Fiora, you would know that her ult would deal less damage if it hits the same target multiple times. Your enemy also has full control of where you land because the first target you pressed R on would be the place where you will end up. Fiora also had no ways to block the CC from your team. So, in order for Fiora to ult you, the 400 range ult, she would have had to Q towards you. Either by flanking from behind or over a wall. Sadly, this wasn't necessary because most adc and mages where playing front line and then complain fiora is OP hahahaahahaha Almost every support was able to interrupt her Q during its animation. And she could only do it twice in quick succession (like in the same 0.5 seconds), then she had to wait for a longer CD than today's haha 2 sec cd on Q. Its also beautiful to see you claim 1000-2000 unavoidable damage, even more "uninterruptible" hahahahaha. Since, as mentioned above, you could've give the "ult" to someone else in your team and avoid a bunch of follow up damage. Locket was also a common build on supports, so, they could've heavily reduce the damage. But yeh, aside from lying about the numbers and calling it more than it was, its also unavoidable hahaha. However, that would require team work, and counterplay, we can't have that. FIORA IS BROKEN OP, OMG SO OP hahahahaha the current fiora can counterplay her enemies more and has less counterplay herself. But, you know, keep telling yourself Fiora was OP just because you people play front line carry haha and farm alone at 25+ minutes. god, this is so much fun. arguing with people that are so uninformed is amusing
You were honestly just waiting to dump your salt on someone, weren't you? Unfortunately, none of that has to do with the price of salt in India. What does the price of salt in India have to do with Fiora? Just as much as your drivel because none of it addressed my points. I also don't remember saying she was OP back then. Her ult was just stupid.
: way to survive original fiora ult: life steal???????????????? you got 4 seconds to do so don't be below 800-1000 hp if fiora managed to get in 400 range of you????????????? buy a ninja tabi????? automatically drops her damage by 10% zhonias if you are ap?????????? spell shield if you are sivir???????? dodge if you are jax?????????? passive if you are phanteon????????? shield if you are a normal adc (have a support near you in mid to late game lol)????????????
I'm going to go out on a limb and say you never played with or against her back then. Because if you did, you'd remember that the bullshit came from her ult procing hydra on every strike and being virtually unhittable for the duration of the ult. Literally nothing you listed would counter her ult any more than just buying a cloth or chain armor. So remember, it wasn't bullshit; it was just about 1000-2000 unavoidable, uninterruptible aoe damage on a champion built to clean up people who are low.
ValSmash (NA)
: What 4800 ip champion should I get?
{{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}} {{champion:117}}
Rioter Comments
Rioter Comments
Rioter Comments
Rioter Comments
: No, its literally every mage mid. Riot has stated that mages burst healing in teamfights is not healthy "in their opinion". In addition, it leads to stale laning where both sides just afk clear and never fall low enough to punish. __________ Problem with Riot stating this is that everyone gets nuked from 100-0 pretty much any time something happens, which means spellvamp would be a bad stat for mages either way.
Honestly, you could probably pull up a Riot post saying what you claimed, and I could find one saying what I claimed because Riot has said both. I'm saying it's more towards the tank draining because while siphoning off minions of assassins resetting health is bad, aoe tanks deal a shit ton of damage to everyone. They're already hard to kill, and unlike assassins who kill one target and then wait a while, the tank just soaks more and more damage while siphoning off everyone in a fight. Plus, if Grievous Wounds is the answer to this, who is more fragile: an assassin or a tank?
: > [{quoted}](name=Z Statistic,realm=NA,application-id=3ErqAdtq,discussion-id=hqMaxEUA,comment-id=00000001,timestamp=2017-02-05T21:16:24.984+0000) > > You obviously didn&#x27;t read anything. > That is the exact opposite of behavior for Spell Haste... "If I had +25% spell haste, then I would be reducing the remaining cooldown by 20% as soon as I got the item" yep that sure wasn't what you said
Yup, and what would happen if you lost 20% spell haste? You'd gain additional cooldown time based on what was remaining. You're still in the mindset of reducing time instead of changing rate of cooldown. If the spell has a cooldown of 90 seconds and after 30 seconds I increase my spell haste to 100%, it would be fully recharged 30 seconds later. However, if after 15 seconds, I sell/get rid of the 100% spell haste, I don't only have 15 seconds left on the recharge. I would have 30 seconds instead. If an item speeds up my cooldown rate and I sell that item, my cooldowns don't keep recharging at that sped up rate. They resume recharging at the normal rate.
Jango Mage (EUNE)
: Why did riot even removed Spellvamp?
Everyone will tell you it's vlad or morde. No. It's Amumu, Sejuani, Zac, etc. 33% effectiveness on aoe spells doesn't really help either.
: you just described the main reason why CDR doesn't work on spells that are already on CD Anyone could just use the ult go to the store, buy CDR and then as soon as it took effect, sell the item back and then buy it again watch the CD drop again and rinse and repeat until the CD on your ult is up. Imagine a Karthus that does this, the instant he reaches lvl 6 he would just be at base buying and selling CDR to use his ult as often as possible making it impossible for the enemy to do anything. Or a Soraka doing the same thing constantly healing her allies
You obviously didn't read anything. That is the exact opposite of behavior for Spell Haste...
Rioter Comments
Rioter Comments
Lakrosin (NA)
: > [{quoted}](name=Z Statistic,realm=NA,application-id=3ErqAdtq,discussion-id=wYIMAxnM,comment-id=00000000000000000000000100010000000000000000,timestamp=2017-01-12T19:04:29.157+0000) > > If I&#x27;m still strawmanning you, then you&#x27;re going to have to start over because the only reason I can think of that you would say Ori&#x27;s is better is because of subjective preference. Nope. You need to read. I said everything I wanted to. You can continue to put words in my mouth. I'm very literal. Never did I say Ori was better than Morgana. However, I do not like Morgana and do feel like she is underpowered compared to most supports.
Now, you're moving the goalposts. You said Morgana's *shield* was weaker than or just as good as Orianna's *shield*. But this is lovely. > [{quoted}](name=Lakrosin,realm=NA,application-id=3ErqAdtq,discussion-id=wYIMAxnM,comment-id=000000000000000000000001000100000000000000000000,timestamp=2017-01-12T20:50:55.451+0000) > > Never did I say Ori was better than Morgana. I didn't say you did. > [{quoted}](name=Lakrosin,realm=NA,application-id=3ErqAdtq,discussion-id=wYIMAxnM,comment-id=000000000000000000000001000100000000000000000000,timestamp=2017-01-12T20:50:55.451+0000) > >However, I do not like Morgana and do feel like she is underpowered compared to most supports. You said you did. First, you say Ori shield > Morg shield. Now you say support Morg is underpowered with the context from the OP that Ori is mediocre if not not sleeper OP. That's what " x is a thing" means after all. So, it's not that Ori is better than Morg, it's that Ori is average or above average and Morg is below average. The entire reason I'm saying this is because I _have_ read your posts. All of them. Multiple times. I do not fire and forget responses. edit: but apparently I'm colorblind and didn't see that you weren't the OP. That changes everything. My bad. {{sticker:zombie-brand-facepalm}}
Show more

Z Statistic

Level 39 (NA)
Lifetime Upvotes
Create a Discussion