Lu: lu little-known: on the other side of the largest online real-time psychological analysis room

cloud network hunting note: speaking of “hero alliance”, people all know that it is the most popular game, but how many people know that it is the largest virtual laboratory? Yes, Riot are players as the experimental object, the study of human behavior. This article take you a glimpse into what, Riot experimental details at the same time, reflect on the network of human emotional manipulation.

once I open black hero alliance, 5 people with teammates bettered, a teammate will angrily said to me: “delete the game, don’t cheating again”. First-person shooter game players could be so bad to ruthlessly execrate others often, but this kind of situation in the hero alliance style lovely game but also common occurance. Although the game has a crocodile fighters, the nine-tailed fox raccoon dog and even a variety of lovely creeps, but could not calm the players the excited mood.

the game picture of fantasy hero alliance and lovely, but the game process often is fierce, many players often easy to emotional mouth don’t choose the words to say some words with racism and sexism. Maybe this kind of rude behavior in the network BBS, and social media can with impunity, but not in hero league games. Riot in Santa Monica, the game developers are adhered to the principle of “encouraging people to” clean up the game, so as to improve the quality of the players. Riot various consolidation measures, such as the player can report report revel, in the form of other players; But more subtle way judged by this system, studied the evaluation technology evaluation machine will identify the player’s behavior, and accordingly give reward.

the hero alliance is the most popular online game, it’s championship game online audiences as much as than the world series and the NBA finals. On the other hand, it is also a huge virtual laboratory, can undertake experiment to tens of thousands of players at the same time, around the clock to collect data across North America, Asia and Europe. Such large data if applied to the study of human behavior, will break through the shackles, cannot be obtained in university laboratory results.

Riot’s struggle to promote the experiment to the positive role of the hero alliance, because the company itself will also benefit in the campaign. Facebook also conducted a similar experiment in 2014, but when the user know company for profit and control their emotions, angrily denounced in succession. After Facebook was not the opportunity to explain his experimental motivation, because the user has regard them as malicious behaviour.

focus on user behavior enterprise more Riot and Facebook. Many companies often regularly launch site investigation, corrective site again according to the survey accordingly, in order to attract visitors to stay longer or more consumption. It is also from the Riot of experiment self-interest (to prevent early exit game), want to attract other people athletics game down new players. Research network game, a psychologist at Jamie Madigan said: “maybe Riot is the only big data was used to track the game company. Its distinctive lies in its scientific experimental methods.”

unlike most companies refuse to promote themselves online experience, how to improve the player’s Riot experiment aim positive public makes it to myself. Riot can open to the public a glimpse into what, understand the company method to improve the quality of the players.

the hero alliance of experiment design by the game designer Jeffrey Lin suggested that he was a doctor of philosophy in neuroscience at the university of Washington in Seattle. At first, his method of study and research of human behavior as well as academic researchers – put 50 to 60 college students in the lab, and then grouping experiments. Now, he led a Riot team players behavior, they will players in the game as the experimental object, social psychology experiment design, and then further to game design, data statistics and scientific research. Riot that appears in the atmosphere in order to improve the quality, the improvement of game players, but it is based on the economic benefits. Because the game free operation mode, in addition to some unnecessary game items, players can free game, so players of the game of satisfaction and a desire to further investment is critical, also is the purpose of the Riot.

Lin’s team had mass using the known results to improve player behavior psychology. Lin said: “one of the most successful is the first time to apply psychology theory. Then try to go away, we come to know observation of behaviors by the players, but the results are often counterproductive.” Since they know it is not feasible to simply copy the classic experiment mode, must break through the bondage of theory laboratory for innovative research.

Riot from the theory of “psychological stimulation”. Existing psychological theory confirms to some stimulus (such as speech, image, or special color), the behavior will subconsciously be affected. Lin’s team also with color testing, want to know the hero alliance five players’ cooperative and individual behavior will be affected by it.

in the “Optimus Experiment” in the study, the news that they will have five players sent to red and blue, and white color as a reference. Play at home in the west, if the information in red warning their bad behavior, often can cause the opposite effect (for example, if someone is pointed out that the errors of his team-mates, they tend to regress), the red warning to make players more in exasperation abuse. And blue information is used to encourage civilized behavior, really effectively reduce the bad behavior.

the results sometimes very unexpected. In the 2013 Game Developers Conference, Lin with data cited an example, impress the audience a smile: a red positive news “who is the most competitive moral player?” But it caused the increase of negative behavior.

in another set of experiments, Lin’s team hopes to further understand the positive behavior and negative behavior in hero alliance spread in the community. Survey results unexpectedly with the “negative bias” in the psychology theory, and negative stimulus or event “” people are more likely to remember the idea.

Lin said: “in the first survey, we found that players tend to focus on the negative news indeed, unintentionally speeded the messages in the network spread in the community. But then we were surprised to find the evidence of positive news spread faster than bad news. It will be our focus this year the object of study.”

although the Riot of experiment the lack of academic psychology theory support, but the large amount of data provided by the game (the daily chat message of about 27 million players and game exchange information), make it can quickly collect data, which make up the lack of accuracy. They can monitor multiple experimental variables at the same time, such as Optimus Experiment Experiment monitoring more than 10 million 217 different variable in the game, while only 10% monitoring in accordance with the plan, also is enough.

Riot of some experiments also has a promoting effect on the game development. For example in every game limitations to malicious message sending players chat mode, significantly improved the behavior of some players. Experienced the limits to chat by other players to report the number of times on average by 20%. And limit chat disciplinary measures than confined players to help improve their quality and effectiveness of the former 4%. Player behavior of change will be for the monthly draw 67 million people online games cause huge change.

Lin’s team statistical data to prove that their experiment results and the effectiveness of disciplinary measures. But I think individual players rarely actively reflect their malicious whether the player has decreased. In addition, I have seen some players frequently report to others, and let the other players to do the same. I think these behaviors will affect the authenticity of the data.

Riot is not the only one large-scale enterprises collect user behavior data. Technology companies such as Amazon, Google and Facebook are often tens of thousands of users to test to himself, that the user response to the current popular network services.

a social psychologist at the university of Virginia Brian Nosek said: “if academic researchers can also be so study on a large scale, research methods and research results will have progress by leaps and bounds.”

Nosek helped develop large data collecting method of social psychology. He graduated from Yale university in 1998, has participated in the early modern network utilization Project to build a virtual laboratory of Implicit web site Project. Project Implicit web site with a series of interesting network test and quiz, effectively helped Nosek and his classmates to survey people in terms of race, gender or sexual prejudice. Nosek recalls: “three days before the amount of data collection is all the data collected in the laboratory than ever before. So, since 16 million time study is conducted on site. Overturned the change of my research methods, received a special psychological effect of a large amount of data it easier.”

at present, the Project of Implicit site weekly visits at about 20000 people, but still can’t compared with Amazon or Facebook users per second. Nosek and other academic researchers, feel sorry for unable to collect personal data. He helplessly sigh: “there is no special channel no access to personal data.”

Riot different performance in it and the cooperation between universities. Only in the last year, the company will develop the cooperation between six and university research projects, including the university of york and London research on hero alliance players name reaction to a person’s character. Another research project at the university of Massachusetts institute of technology, it is to measure five new players between the team cooperation ability and ability to develop a “collective intelligence” (can be predicted that perform a specific task) test plan.

the game psychology expert Madigan says: “Riot is the only one to study such a transparent game company, even as a marketing tool, like, ‘we tried to improve the gaming community, let you get the best gaming experience. ‘slogan to attract new players.”

the public to avoid the Riot was not informed by the experiment of the player’s complaints. Madigan said: “most of their research is around, on improving the quality of the players who will be strongly opposed to improve their own quality?”

, by contrast, Facebook and Cornell university study is a tragedy, they study of “emotional contagion” PNAS disclosed in 2014. Facebook had not publicly will experiment, so the user informed after angry, because the experiment in order to avoid users affected by friends published by the emotional message, distort the user’s messages. Users lambasted network control their emotions. People questioned why the experiment under the condition of without seeking the opinions of their conduct, whether it passed the relevant examination institutions for the study of ethics of the audit.

in fact, the Cornell censors did review in the study, and unwittingly passed the study. Because Facebook maverick to distort the user a message, and the data collected before to research at Cornell university experimental results. Icahn university joint graduate school of medicine in mount sinai bioethics Michelle Meyer, director of the bioethics policy plan to write: “no ethical audit institutions agreed to the ‘subjects’ research on the user and Facebook stubbornly carried out the experiment.”

Lin emphasized Riot team review, attached great importance to the university audit institutions must pass the examination after approval and academic researchers to carry out the “hero alliance” research cooperation. This commendable attitude will help between companies and universities through ethical review of research work. Future research may accept a higher standard of moral supervision and public disclosure of company.