The Great Trollbunal Experiment

123
Comment below rating threshold, click here to show it.

The One Soraka

Senior Member

12-28-2012

So, seeing as I don't give a damn about my 'justice rating' or wtfever it's supposed to be, I decided to conduct a little experiment. I will be participating in the trollbunal... not as a judge or anything, no no no, but as someone who counts 21 seconds, then left clicks my mouse, then waits 21 seconds, and left clicks again, 20 times a day.


The purpose of this? Well there's always a lot of false reporting going on, and back when I actually did care about the trollbunal, I would often see cases where someone is reported for intentionally feeding, when they went 3/9/8, while the rest of their team has scores like 0/5/2, 1/15/8, 5/6/0, 4/4/1.



So yeah, i'm going to do this for a couple weeks, and see what happens. If this were a working system, i'd imagine my success rating would be somewhere between 40-60%, but knowing what I know about the place, I imagine it will actually be something above 90%, which would indicate something being pretty broken.



Anyone care to make any bets on what 2 weeks of 'punish everything' would get as far as accuracy goes?


Comment below rating threshold, click here to show it.

Harate

Senior Member

12-28-2012

Quote:
Originally Posted by The One Soraka View Post
So, seeing as I don't give a damn about my 'justice rating' or wtfever it's supposed to be, I decided to conduct a little experiment. I will be participating in the trollbunal... not as a judge or anything, no no no, but as someone who counts 21 seconds, then left clicks my mouse, then waits 21 seconds, and left clicks again, 20 times a day.


The purpose of this? Well there's always a lot of false reporting going on, and back when I actually did care about the trollbunal, I would often see cases where someone is reported for intentionally feeding, when they went 3/9/8, while the rest of their team has scores like 0/5/2, 1/15/8, 5/6/0, 4/4/1.



So yeah, i'm going to do this for a couple weeks, and see what happens. If this were a working system, i'd imagine my success rating would be somewhere between 40-60%, but knowing what I know about the place, I imagine it will actually be something above 90%, which would indicate something being pretty broken.



Anyone care to make any bets on what 2 weeks of 'punish everything' would get as far as accuracy goes?
Oh that's crazy I thought it was 30 seconds. I wonder why it starts at 21


Comment below rating threshold, click here to show it.

Explodeurmum

Junior Member

12-28-2012

Your accuracy rate will be about 80%.


Comment below rating threshold, click here to show it.

The One Soraka

Senior Member

12-28-2012

Quote:
Originally Posted by Harate View Post
Oh that's crazy I thought it was 30 seconds. I wonder why it starts at 21
Well, it's 20, but when it counts down, '0' is a whole second, so i need to wait exactly 21 seconds between cases.


I'm also going to take this time to point out, that when I really did the trollbunal seriously, when I carefully read every report and every chatlog, oftentimes taking 5-15 minutes on each case, they had a captcha to prevent spambotting...



...nowadays, there's no captcha, in fact, there's even a SECOND punish/pardon button at the bottom of the page, instead of one JUST at the top? Why would this be?


My guess is sometimes it'll take someone 20 seconds to scrollwheel down through the chat log, and then they'll see the scores, and decide right then and there, punish or pardon, based on the scores, and oh look, here's an extra convenient button with which to do so...


/golfclap


Comment below rating threshold, click here to show it.

Sereg Anfaug

This user has referred a friend to League of Legends, click for more information

Senior Member

12-28-2012

If you believe it's broken, why continue to break it? If you're going to spam something, and you actually don't care about your rating, wouldn't spam pardon be more your style?

In any case, an 80% Punish rate is far from broken - people who are reported enough to make it to Tribunal are very likely to be guilty. False reports do happen - which is why people can even be Pardoned. But there's no reason to believe that a system specifically designed to remove the most toxic members of the community would have a Punish rate of only 50%.

As far as intentionally feeding reports - those are usually the wrong category. Actual intentional feeding is quite rare. However, most of the cases I come across with incorrect intentional feeding reports are still guilty of another offense, usually verbal harassment, offensive language, and/or negative attitude, and as such, should still be Punished.

This isn't the real world, where the punishment matches the offence. This is League of Legends, where the punishment corresponds to how many times you've ignored previous punishments.

So you don't have to worry about a Punish for a reason other then the report reason - it'll work out, trust me =P


Comment below rating threshold, click here to show it.

gnfnrf

Senior Member

12-28-2012

The Tribunal punish rate is around 75% to 80% right now, so that's about what you'll get. Several other people have tried this, and they've all gotten around there. I've analyzed 10,000 cases from reform cards and gotten 76%, and someone else did 14,000 cases and got something around there as well.

Just so you know, after a few days of this behavior, the Tribunal will detect that you are spam punishing and stop counting your votes towards the actual outcome. So, you won't be gathering new information, and you won't be taking it to the man and corrupting the system. But you can feel free to click the button as much as you want.

Of course, an 80% conviction rate doesn't necessarily imply anything about how accurate the Tribunal is. If Riot is clever enough that 80% of the people they build Tribunal cases for in the first place are guilty, then the Tribunal could be 100% accurate (of course, that is a practical impossibility).


Comment below rating threshold, click here to show it.

JesterTheJoker

Senior Member

12-28-2012

Quote:
Of course, an 80% conviction rate doesn't necessarily imply anything about how accurate the Tribunal is. If Riot is clever enough that 80% of the people they build Tribunal cases for in the first place are guilty
Pretty much this... Not everyone gets reported equal amounts, and although there are obviously false reports at times, its also clear that most people that end up in tribunal are guilty (from the cases I've reviewed). Out of the 35 cases I've reviewed, I've pardoned 4 or 5, and in fact the only case that I didn't agree with the majority was one where the majority pardoned, and I wouldn't have since he was running about calling everyone r-tards, idiots and mother-f'ers, but hey thats just me.. i recognize that the system is what it is.. Community moderation of behavior. No system is going to be 100%, but to take incrediably small minority of cases where a mistake is made out of the tens or even hundreds of thousands of cases and call that proof the system doesn't work is idiotic.


Comment below rating threshold, click here to show it.

Great Pyrenees

Senior Member

12-28-2012

So your going to troll the trubnal by punishing every case? +1 upvote for you sir


Comment below rating threshold, click here to show it.

Fudouri

Senior Member

12-28-2012

Just want to second the idea that if you feel tribunal is overly harsh, you should be spamming the PARDON button and not the punish.

Also, punish rates are just under 80% right now so that's what you will get if you spam punish, no real point in actually doing it since we already know the results.


Comment below rating threshold, click here to show it.

magooomba

Senior Member

12-28-2012

Quote:
Originally Posted by The One Soraka View Post
If this were a working system, i'd imagine my success rating would be somewhere between 40-60%, but knowing what I know about the place, I imagine it will actually be something above 90%, which would indicate something being pretty broken.
In case this point hasn't already been hammered home, allow me to delve into this. You say that the Tribunal would be seemingly "pretty broken" if the punishment rate would be above what your imagined success would be. That is a ridiculous conclusion based solely on your own assumptions. If X% of the players who enter the Tribunal are "deserving" of punishment, simply claiming that X% of players being punished somehow shows a supposed problem is a joke. A punishment rate alone shows nothing but how many players are being punished.


123