Statistical Variance and How it Impacts Matchmaking

Randomonium··1 votes·568 views

Elo hell. Its a phrase that is commonly thrown around but people really don't understand how or why it happens. The answer is statistical variance. A common counter argument to the existence of elo hell is that you have a better chance getting a troll on the other team than you have getting a troll on your team (since you are guaranteed one player, you, isn't going to troll on your team.) While this makes sense on the surface it really shows a lack of understanding about statistics.

For simplicity sake lets say you have an equal chance of getting a troll on your team or the other team. Essentially it turns into a coin flip scenario. Heads you get a troll, tails the other team gets a troll. In theory you should get an equal number of heads as tails but we all know from experience that isn't the case. When you expand your sample size to millions of players even unlikely scenarios like flipping 20+ heads in a row become very probable. This is how some players seem to get trolls every game and wind up in "elo hell."

A common response to this is "just play more games, it'll even out eventually." Again, in theory this sounds great but it doesn't make much sense from a statistical standpoint. Just because you flip heads 20 times in a row does not mean you have a higher chance of getting a tails on the 21st flip. Each flip is always a 50/50 chance. It is true that if you flip the coin an infinite number of times you will eventually break even but because very few people play more than a couple hundred games a season it is very possible that they will never regain the elo they lost.

The point of all this is to not justify why I or any other person is in bronze, only to better educate the community that these strings of "bad luck" are not only possible but probable for some players. And while Riot tries to "weight the coin" with its matchmaking system, even a fantastic matchmaking system (which I am not saying Riot has) cannot get rid of statistical variance entirely.

So what can you do if you are one of these statistical anomalies? Weight the coin as heavily in favor of yourself as you can. Try to duo with people you trust. Communicate positively with your team and give compliments when they get kills. Invest heavily in wards and encourage others to do the same. Avoid 50/50 scenarios where you are just as likely to get killed as the enemy (i.e. only engage when you are sure you can get the kill.) Encourage people you play with to join a Curse call so you can communicate better. If your team is behind avoid engaging the other team in open ground until you catch up in gold (i.e. split push, bait tower dives, etc.) If the enemy team gains control of an objective don't go into a bad team fight, push another objective instead (i.e. trade dragon for a tower or trade one tower for another.)

If all else fails try to remember its just a game and find other ways besides soloQ to have fun. I stopped playing soloQ for months and instead focused on playing with teams and in tournaments and I am enjoying LoL way more now than the soloQ grind.

2 Comments

Randomonium1 votes

Hmm, I guess math really does put the kids to sleep these days.

Ice Weasel X1 votes

{quoted}

Hmm, I guess math really does put the kids to sleep these days.

Try pictures. Even in stuffy corporate meetings, visual aids help break the monotony. People see text walls and stop short. Also, length can be daunting (...ladies). If it looks like it's gonna take time to read, some people will assume it's boring. Or even if it's not boring, just too much of a time killer. One of my friends enjoys auditing MIT lectures while he works, but will only listen to them at double speed playback because he's too impatient to wait for the information at normal speed.

More from Gameplay

More by Randomonium