Why the "it will all even out eventually" is a bad argument for trolls and afks losing lp's

slippykitten·Oct 7, 2015·162 votes·9,104 views

Riots official statement on why they don't do anything to compensate players for having an afk or being trolled is that eventually the other team will have the same scenario and it will all end up evenly.

**This is not only lazy, but it is bad math. **

Let me use a coin for an example, if you flip a coin x amount of times, you would expect to have an even number of heads and tails, but you wont. Do it, pick any number of flips greater than 10 randomly and see if you get the same number of heads as you do tails. You probably won't. If you kept flipping infinitely, you would eventually hit an even number, but it would likely take thousands of flips. This is because the probability of getting a heads or tails is 50/50 no matter how many times you flipped. If you get heads 1,000 times in a row, your next flip still has a 50/50 shot of being heads, and there is no guarantee that this will even out in a reasonable amount of time.

This is exponentially more problematic when you stick 10 variables into the equation instead of two. Setting aside all other factors (which are already vast) and only looking at the probability of having a troll or afk on your team, it is statistically unreasonable to say that you will get exactly the same amount of trolls or afks on your teams as on the enemy teams in a reasonable amount of games. Without doing all the math here, it would take tens of thousands of games to feel stastifically confident that each team would have an even amount of afk's or trolls.

This means that some players are unfairly benefiting from this and others are unfairly hindered by it. This problem creates other problems, as if it takes longer for players to reach their actual rank, then they are unfairly impacting the games they are in, creating scenarios where other players win or lose games they should not have.

**At the end of the day, this means that matchmaking is a crapshoot. **

tl;dr - For a person of even reasonable intelligence to suggest that it's ok to have afk's and trolls, because you are statistically likely to even out in the end, is intentionally dishonest. Essentially, Riot is intentionally deceiving their playerbase to avoid addressing gamebreaking problems, and that isn't ok.

Edit A lot of people have been pointing out that you are more likely to have the troll on the enemy team than you're own. I'll point out two important points on that:

That's only true in an extremely large sample pool. Once again, this is the gamblers fallacy. Yes, if you are not afk or trolling, then over 1000's of games, you will have had more on the enemy team than you're own, but it is still statistically possible to have 100 consecutive games where you were the one trolled or afked on.

and much more importantly 2. This doesnt make it ok. You are only pointing out how trolls and afk's may artificially inflate someones elo, which is still BAD! This still negatively impacts players, makes the game less fun, and destroys matchmaking.

**Trolls and afk's will never "even out" for the average player to where their influence was nil, and it is not ok to do nothing about them. **

Edit 2

A very much related problem - Riot has responded to over 50 posts that have less people posting on them as this one. When a company cares more about their PR then their actual players, it's hard to believe they are doing something about it.

Here's some thoughts.

there are 3 factors that disagree with the proposed notion.

Law of large numbers Which basically says what you see in a small number of games, doesn't necessarily reflect what you will see in a lot of game. You can get an AFK/Troll 20 games in a row, but still have it balance out over the course of 100-1000 games.

Basic laws of probability Assuming you're not a troll and 1 per game, you have a 4/9 chance of getting the afk/troll on your team; 5/9 of the enemy getting it. So odds are better that you benefit from the trolls than you suffer.

Repeated proof of advancement Yes, I'm going to poke the pink elephant. It doesn't matter how many trolls are in your normal matches, if your good enough, you'll climb. Statistics, streamers, and Even Riot's experiments using pros have proven this. So whether you have a lot of matches or a few, is irrelevant. The system is proven to work.

Then there's 2 aspects that can be used in support of your claim.

Toxicity contamination People are more inclined to remember negative experiences than positive ones. Lyte even did a community experiment back on the old forums that demonstrated this. When people experience toxicity, they're also more likely to become toxic themselves. This means the chances of experiencing negative behavior (which includes AFKs/Trolls) increases exponentially the more you play during a given session. And people contaminated by toxicity are also more like to perceive toxicity. This causes otherwise isolated events (Amumu missed the smite steal) to appear as something much worse (Amumu intentionally feeding). Of course, the toxicity equation gets thrown for all sorts of loops when you throw in rebuffing positive players. But we're getting into an entirely different topic at that point.

elo Pitchforks down! Riot has acknowledged that the density of negative behavior goes down as elo goes up. (It never goes away.) So if we assume the density is enough for 1 per game, and 1000 games are played over 10 elo brackets, then we can reasonably assume that the density of negative behavior at the top elo bracket is half the bottom (exact value is irrelevant right now). Meaning that the top 100 games will see only 50 instances of negative behavior while the bottom 100 would absorb the remaining. The law of large numbers still works if you look at individual brackets or all brackets together, but falls apart if you factor the brackets against each other.

PROMOS It's already well established that promos aren't like normal matches. The system intentionally stacks the odds against you to see if you genuinely deserve to advance. This means [in theory] the enemy is better and/or your allies are worse. Now remember, the system determines better/worse through elo. Higher elo has a lower chance of negative behavior. Which means, even if it is by a fractionally small amount, your odds of encountering negative behavior during promos is [theoretically] higher.

I'm not posting this in support of either camp. They're both correct and incorrect depending on how you look at the information given. But having said that, we do encounter the giant monkey wrench in this entire debate. Most of our information comes from unreliable sources.

Personal Experience People are horrible at being unbiased - especially about them selves. Hitler thought himself a servant and hero; the uni-bomber thought scientists where out to get him. You think you're under siege by toxic players, I think there are ninjas waiting in my bathroom to abduct me. Horrible. Absolutely horrible.

Riot. But Riot holds all the raw data, they can't be unreliable! WRONG! Reliable sources can be identified, reviewed, and tested. For everything Riot tells us we can............ Only identify it...... Sometimes. Now, I'm not saying they're wrong. They have better access to that raw data than anyone; which makes them the most knowledgeable source for that information. But because they don't openly share that data or it's collection methods, we can't review or test it. Basically, The worlds best spiritualist is saying there is a ghost behind me. But they can't prove it to me because it would scare away the ghost. Also, they're saying my neighbor has weapons of mass destruction. (>_>)

Anyways. Just a bit of food for thought.

DISCUSS!

Why the "it will all even out eventually" is a bad argument for trolls and afks losing lp's

173 Comments

Here's some thoughts.

More from Gameplay

More by slippykitten