Dear everybody at Riot:

First Riot Post
Comment below rating threshold, click here to show it.

Tarrent

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

Junior Member

11-13-2012

Quote:
Originally Posted by Lyte View Post
7 No one implies a causal relationship anywhere and numerous psychologists reviewed the language to double check the nuances of the tips.
Lyte, my girlfriend has a masters in psychology and loves League of Legends and wants to know how she can get this job, haha.


Comment below rating threshold, click here to show it.

Dracorya

Senior Member

11-13-2012

Quote:
Originally Posted by Tbird View Post
That just happened
+1 Sums up exactly how I feel every time these people try to claim data is "wrong" when in actuality they have no idea what they are talking about.


Comment below rating threshold, click here to show it.

Lyte

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

Lead Social Systems Designer

Follow RiotLyte on Twitter

11-13-2012
3 of 4 Riot Posts

Quote:
Originally Posted by Tarrent View Post
Lyte, my girlfriend has a masters in psychology and loves League of Legends and wants to know how she can get this job, haha.
davin heads up a research team at Riot in the Player Experience department and is looking for 1-2 more researchers. May want to look into that as I don't know the exact job description off-hand.

On the player behavior team we are not actively searching for new members but we're always open to exploring opportunities and creating brand new positions for exceptional people.


Comment below rating threshold, click here to show it.

Knife Kitty GG

Senior Member

11-13-2012

^ ;p


Comment below rating threshold, click here to show it.

davin

Senior User Researcher

11-13-2012
4 of 4 Riot Posts

Quote:
Originally Posted by Lyte View Post
davin heads up a research team at Riot in the Player Experience department and is looking for 1-2 more researchers. May want to look into that as I don't know the exact job description off-hand.
This one: http://www.riotgames.com/careers/ass...ser-researcher

We'd love to have another qualified, passionate researcher join us here


Comment below rating threshold, click here to show it.

Ladayen

Senior Member

11-13-2012

heheh you said ass(ser) researcher)

Sorry massive beavis and butthead flashback there lol.


Comment below rating threshold, click here to show it.

LittleDi

Senior Member

11-13-2012

Quote:
Originally Posted by Lyte View Post
There's a difference between inability to understand data and releasing misleading data. For example, everyone who understands data knows that the in-game tips are correlations and not a causal inference. No one implies a causal relationship anywhere and numerous psychologists reviewed the language to double check the nuances of the tips.

If you want to discuss specific examples, let's do that. Otherwise, you aren't adding anything to these boards.
Still think that total of false punishes out of how many cases total wouldn't hurt to be known once if ever. I already know it happened at least once. But if PB&J could actually sit down, get the number of overturned cases due to complete false punish verdicts out of how many exist in cases It'd be an extremely useful statistic in telling people how rare they really are.

I'll be very specific then.

Quote:
Originally Posted by Lyte View Post
2) What actions did Riot take?
a) In this small group of players, we decided to make their reports have 0 power. In addition, we looked for players who were in this group and also recently displayed toxic behavior in their games or had a Tribunal history and gave them an additional 7-day ban. This was the worst 0.2% of these players which ended up being a few thousand players per territory.
This is not too terribly in the past. This was taking from the most toxic of the toxic and getting a very small target area of this. It aimed for the absolute worst end of the spectrum. What I am asking for is more of a challenge. Pointing out the absolute best of the worst whom actually might not of been in the toxic category at all.

Chances are that many of these have factors that will not make sense if we dissect each case by case. Which is why I am more concerned for those who are well above this as well. So make a study for yourself to complete PB&J. If we go from looking at the 98.6% (number may of changed) to looking at the 98.7% followed by finding the least offensive 1% of that how many false punishes would that be discovering?

Narrow it down as much as possible.
Barely reported players who just squeaked into the tribunal's radar would be a starting point. Chop that down by going after first time warnings. Cut that even further down by finding players who've been around the longest without reports. Cut that by total of reports per game so we're looking at single report case files that would be loaded by the tribunal.

Is that a small enough percentage to determine how many possible false punish reports would exist? I would be considering that the closest I could get without narrowing it down further without your end data. So I'll stack onto all of this anyway. Get at least 2 more things I wouldn't ever know about because I don't work for Riot and add those to the test factor. Go for broke if you must to get it down even further than that so you're looking at a total of 100 cases or so out of the highest end spectrum of the tribunal. If you can get it down to 50 from that. Look these cases over, but only after they cycle thru the tribunal.

Consider it a test to see if Lawrence Berkeley National Labs most powerful microscope could find anything at this extremely small of case file. Then without releasing exactly what kind of numbers you're looking at get the percentage of chances of false punishes out of this group. You could go with 100, or 1000, or 10k but I figure the smallest possible case size would give you an accurate reading of how insanely difficult it would be to get that barely into the tribunal and get false punished.

Please note I'm not asking Riot to judge them. Just find the least likely offenders bracket, send them thru the tribunal and come back with a result of how many false punishes actually happened. All this would confirm or deny would be how accurate the tribunal is. Knowing that you're not actually telling us how many cases cycled thru with them force this into a percentage post much like the one you did in the section I quoted above.

We already know the tribunal is very accurate at taking out the worst of the worst offenders. So study how accurate it is on the other end of the spectrum and give us an update from what you could.

Best case scenario: You need to tighten up the process as it's picking out people who really shouldn't even fall into tribunal in the first place.

Worst case scenario: Tribunal needs more reach and the minimums with which we see EVERYBODY is beyond guilty with not a single false punish possible.

Since nobody outside of Riot could even say how small a window this is aiming at (possibly the least offensives 0.001% section) everybody could guess whatever their heart's content would be and still be clueless of where exactly that line is due to this study making it change anyways. There probably already are some people trying to play that hole in one already. This would just move the green regardless of how accurate they were.

This is about all I could figure out on my side of the monitor. So... How about it Lyte? Is this a small enough window to study?


Comment below rating threshold, click here to show it.

Brackstone

Member

11-13-2012

Quote:
Originally Posted by Lyte View Post
There's a difference between inability to understand data and releasing misleading data. For example, everyone who understands data knows that the in-game tips are correlations and not a causal inference. No one implies a causal relationship anywhere and numerous psychologists reviewed the language to double check the nuances of the tips.

If you want to discuss specific examples, let's do that. Otherwise, you aren't adding anything to these boards.
The Lyte! It BURNS!!!!!

On a more serious note I'm glad to see you standing so firmly by your data and your numbers. It speaks not only to your character but more importantly to how seriously you take your job and I hope it gives you a lot more credibility in the community. Too often I feel like (no matter what game I'm playing) developers or staff members always seem to waiver in the face of their customers criticism.

+1


Comment below rating threshold, click here to show it.

ShadyElf

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

Senior Member

11-13-2012

Quote:
Originally Posted by davin View Post
This one: http://www.riotgames.com/careers/ass...ser-researcher

We'd love to have another qualified, passionate researcher join us here

you guys might wanna tone down on the fire on that page


Comment below rating threshold, click here to show it.

LittleDi

Senior Member

11-13-2012

Quote:
Originally Posted by ShadyElf View Post
you guys might wanna tone down on the fire on that page
They should send in Firefighter Tristana to take care of some of that.

PS: http://www.riotgames.com/themes/riot...careers_bg.jpg
It's an annie picture.