Boards›Player Behavior

Punishment

Awf Meta·3/5/2018, 11:41:00 PM·1 votes·541 views

http://open.lib.umn.edu/intropsyc/chapter/7-2-changing-behavior-through-reinforcement-and-punishment-operant-conditioning/ The link above is just so people can get familiar with the terminology and general knowledge.

The link below, points out the part I want to specifically discuss. https://quizlet.com/121225710/psychology-chapter-6-learning-flash-cards/

Factors that influence the effectiveness punishment -timing -intensity -consistency Timing is the most effective when applied during the misbehavior or as soon afterwards

Intensity -punishment has to fit the crime -too mild=no effect -severe punishment=adhesive side effects

consistency -must be applied consistently -parents need to be in the same page/react the same

Problems with Punishment -doesnt teach/promote alternative acceptable behavior -may produce undesirable results/hostility/fear/passivity -may be temporary -can model aggression

Now let's consider League.

Timing - Some accounts may exhibit an unwanted behavior for hundreds of games. Best case scenario, IFS triggers a punishment minutes after a game. Still considered delayed, because punishments are handed out after a game not after the offense.

Intensity - All offenses can be punished with the same intensity. This is the tier system. While mild offenses are slower up the tier system, ultimately, they receive the same punishment as severe offenses. One would have to argue repeat mild offenses are equal to severe offenses.

Consistency - I'm a little split on this one. Every instance is not always punished. Consider intentional feeding cases. It usually takes a history of offenses before punishment is issued. I would say, well defined offenses are consistently punished. The more subjective offenses are not consistently punished.

The problems with punishment in League context:

Teaching in game play context - there is almost no teaching. It is very hard to show how game play would be different with alternative actions. There is a lack of "teachers".

Teaching in chat context - There is almost no teaching for positive chat. There is teaching for muting or neutral chat.

Undesirable results - I think there is an argument to be made. Some punished players turn to "troll accounts". Some punished players become afraid to use chat all together.

May be temporary - I think we see this a lot. I don't have actual data though. Keep in mind, mild offenses usually require a threshold before being punished. It may take some time before the temporary state of reform is discovered. The punishment it self is pseudo temporary because of the ability to create new accounts.

Can model aggression - That is a tough one. I feel Riot shows little empathy. I'm not entirely sure that constitutes being aggressive. EDIT: The modeling of aggression shows in the communities unwillingness to come to peaceful resolutions. The community mutes and reports the instant they see something objectionable. Furthermore, they desire the offender to be banned without question.

18 Comments

Chermorg3/6/2018, 12:10:29 AM7 votes

Teaching in game play context - there is almost no teaching. It is very hard to show how game play would be different with alternative actions. There is a lack of "teachers".

I would say that there's not really a need for this - Riot gives tips already that have the statistics on how positive players are more likely to win games. Riot cannot magically make someone behave appropriately and then they see they're winning more games - the only thing they could do is make a dev-blog about this.

Teaching in chat context - There is almost no teaching for positive chat. There is teaching for muting or neutral chat.

There's the Summoner's Code, or is that not what you're referring to?

Undesirable results - I think there is an argument to be made. Some punished players turn to "troll accounts". Some punished players become afraid to use chat all together.

The number of players who turn to "troll accounts" is lower than you'd expect, but I will accept the truth that it does happen. Are the very few people who turn to trolling worth the hundreds who are punished for toxicity for every one or two troll accounts that are made? I personally think it is worth it because then you have many more positive players than trolls.

As for players being afraid to use chat, I certainly don't experience this. I think players may be afraid to the point they'll mute people before responding to them.. but that's kinda the goal.

May be temporary - I think we see this a lot. I don't have actual data though. Keep in mind, mild offenses usually require a threshold before being punished. It may take some time before the temporary state of reform is discovered. The punishment it self is pseudo temporary because of the ability to create new accounts.

I think that the point in your quote is referring to the improvement after a punishment - and I agree, it can be a temporary improvement and then a return to poor behavior. This is why Riot has escalating punishments, but also allows accounts to demonstrate good behavior for a while to "drop down" and get a chat restriction "warning" again instead of a higher punishment.

Can model aggression - That is a tough one. I feel Riot shows little empathy. I'm not entirely sure that constitutes being aggressive.

I don't think Riot shows "little empathy". Rioters play the game - they experience trolls, they experience feeders, flamers, etc. They understand what it feels like. Those feelings do not justify toxicity, and Riot does not need to remove/lessen punishments to show that they're empathizing.

I'd like to talk about the characteristics of the punishment now:

Timing - Some accounts may exhibit an unwanted behavior for hundreds of games. Best case scenario, IFS triggers a punishment minutes after a game. Still considered delayed, because punishments are handed out after a game not after the offense.

While this is true, punishments aren't issued after any game - they are issued within 15-20 minutes after a poor game, meaning that at most one game passes between an exhibition of poor behavior and a punishment. The timing of the system could not really be made much better without compromising the analysis and determination the system makes as to if something is punish-worthy or not.

Intensity - All offenses can be punished with the same intensity. This is the tier system. While mild offenses are slower up the tier system, ultimately, they receive the same punishment as severe offenses. One would have to argue repeat mild offenses are equal to severe offenses.

I think that it is very arguable that repeated mild offenses cause just as much damage to the community as one severe offense - if not more. If someone uses hate speech in a game (resulting in an instant 14 day ban) - they affect 9 players. However, someone displaying minor toxicity over 9+ games (the number required for the average escalation to a 14 day ban from nothing), they're potentially affecting 80+ players. The end result is that while 80+ players may not have had to deal with racism/sexism/etc, they still had their game experience detracted from.

Consistency - I'm a little split on this one. Every instance is not always punished. Consider intentional feeding cases. It usually takes a history of offenses before punishment is issued. I would say, well defined offenses are consistently punished. The more subjective offenses are not consistently punished.

The issue with intentional feeding is that Riot does not issue a punishment until they are certain that an account was intentionally feeding - be this via the automated system (colloquially called FeederBuster), or via a human review (either by itself or in combination with a review of a text punishment). It's as if the parents only had circumstantial evidence that the kid stole a cookie - but in fact it was the dog. Should the parents punish the kid, even if they find 10 cookies missing, but they have no proof that it was the kid? They may have thought they heard the kid in the kitchen, they may have seen him walk down the stairs, they may have seen the jar on the counter - but is there any proof it was the kid?

I argue that this is an example of just how consistent Riot is - they will not punish an account until it is beyond any reasonable doubt intentionally feeding. The system treats all accounts the same - once it goes beyond doubt that they're intentionally feeding, they get slapped with an instant 14 day ban.

ModPeriscope3/5/2018, 11:52:03 PM6 votes

I have no idea what you're writing about. There's ideas, sure, but none of them focus around anything other than vaguely referencing the punishment system.