So overall after all the blah blah we found that THE ELO SYSTEM IS NOT AN ACCURATE REPRESENTATION OF SKILL LEVEL!

We took a sample of 40 highly rated players throughout the world. We used the League of Legends ladders database to select samples. After assigning numbers to the group, we generated random numbers 1-10, to determine which page we would get the sample from. We then generated a second random number using our calculator to determine which player on the page we would sample. After selecting our samples, we used lolking.net to look up their recent win proportions during their last 10 games.
The purpose of the project was to determine whether highly rated players win more than 50% of the time. We intend to perform a 1 proportion z test to determine if high rated players win more than 50% of the time. If they do, than the elo ranking system is justified for awarding the high rating to winners. They key issue is whether or not the elo system of ranking is effective or not. The null hypothesis is Ho=.5 and the alternative being Ha>.5. In the terms of a simpleton, we are testing to see if a person labeled as skilled actually deserves that label.

After we collected all the data...

__________
p=.5625 sd(of sample so p hat) = \l (.5)(.5)/40 = .79
q=.4375
conditions:
SRS
population>10n 10x40=400, more than 400 players total
np>10, 40x.5625 = 22.5
nq>10, 40x.4375 = 17.5
We used a 1 proportion z test because we are using the proportion of wins for each sample from their last 10 games played.
Ho: Probability that highly rated players win as much as they lose Ho=.5
Ha: More highly rated players will win more than they lose Ha>.5
Test Statistic: observed-predicted/standard deviation
(.5625-.5)/.79=.07911

Because the proportion was only .5625, we fail to reject the null hypothesis.

As you can see, THERE IS NO EVIDeNCE TO SUGGEST PLAYERS WITH HIGHER ELO ARE BETTER!
yay teemo

Increase your sample size and do it again.

So basically.

What you're saying is that a random tiny selection of higher elo players have an average/upper win rate in games of their skill.

If you place a high Elo player in low Elo enviroment, they will win more then they lose.

What you did, was take high Elo player who were in a high Elo enviroment and predicted that they win as much as they lose.

Elo working as intended.

op failed at stats

well.. not at using formulas, but applying the results of a test. drawing a conclusion. designing a test. you know, pretty much everything that matters besides calculator functions.

I gotta ask why all the hate in regards to this post?

At least it's an attempt to provide some sort of justifiable reason with supportive evidence why the Elo system doesn't work.

Regardless of whether or not you agree with the findings at least the argument has more merit than most posts on Elo.