League of Legends honor statististics

12
Comment below rating threshold, click here to show it.

iatre

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

Senior Member

12-16-2012

I was wondering if there is any correlation between the number of honor badges on a team and the likelihood that the team wins, so I've started collecting data. I realized that this would take an incredibly long time if I were doing it myself (plus the numbers will be a little skewed, since I'll be in every game and I have a badge). So, I've started a Google docs spreadsheet with a submission form for all to use:

https://docs.google.com/spreadsheet/...ekktMzliUVE6MQ

Using this form is very simple. Just select how many badges your team had (team 1), how many your opponents had (team 2), and who won the game.

If you're playing a ranked game, this gives you something to do during champion select. Just open the form and select the numbers of badges, then after the game, select the winner. Then, you have the option of entering your summoner name, just to keep track of who sends the most reports.

For now, this is for 5v5 summoner's rift, not against bots. If this goes well, I might expand this to include other data points.

Once I get enough data, I'll post my results. Thanks!


Comment below rating threshold, click here to show it.

iatre

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

Senior Member

12-16-2012

Here is a link to the current submissions. At some point, I'll modify this to include some statistical analysis.

https://docs.google.com/spreadsheet/...tMzliUVE#gid=0


Comment below rating threshold, click here to show it.

Will Oak

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

Senior Member

12-16-2012

Bumping and upvoting for support


Comment below rating threshold, click here to show it.

ninjamast3r

Senior Member

12-16-2012

Hmmm, looks interesting. If I remember next time I play, I will help.


Comment below rating threshold, click here to show it.

iatre

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

Senior Member

12-16-2012

Quote:
Originally Posted by ninjamast3r View Post
Hmmm, looks interesting. If I remember next time I play, I will help.
If you bookmark the submission link, it might help you to remember.


Comment below rating threshold, click here to show it.

iatre

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

Senior Member

12-16-2012

I estimate I'd need a couple thousand games for statistical significance, so I hope some people help me out with this project by submitting games.

Does anybody know of any other way to get honor badge info? I don't see a way to get this from API.


Comment below rating threshold, click here to show it.

Suave91

Member

12-16-2012

Quick tip: you might want it "team 1" to be blue team and "team 2" purple team. A lot of people claim Summoners Rift unbalanced in blue's favor and you can cut out a confounding variable this way.

Additionally, if you did a "which color won" option as well, you could do a Chi-square test comparing honorable teams and colors to see if honor has a statistically significant impact. It would solve two questions with one survey.


Comment below rating threshold, click here to show it.

iatre

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

Senior Member

12-16-2012

Quote:
Originally Posted by Suave91 View Post
Quick tip: you might want it "team 1" to be blue team and "team 2" purple team. A lot of people claim Summoners Rift unbalanced in blue's favor and you can cut out a confounding variable this way.

Additionally, if you did a "which color won" option as well, you could do a Chi-square test comparing honorable teams and colors to see if honor has a statistically significant impact. It would solve two questions with one survey.
It's an interesting thought, but I don't think that blue/purple really makes a difference for this survey. One would expect that there'd be an equal number of blue teams and purple teams with more honor badges, so any effect should cancel out. I mean, there's all sorts of variables I could try to account for if I wanted to be thorough, but my key assumption is that as the number of games gets very large, other factors should cancel out (like, number of 4v5 games, number of duo queues, etc).

Of course, my background is in physics, not statistics, so I could be wrong there.

My goal was to keep the submission form as simple as possible to make it easy for people to participate.


Comment below rating threshold, click here to show it.

Suave91

Member

12-16-2012

Quote:
Originally Posted by iatre View Post
It's an interesting thought, but I don't think that blue/purple really makes a difference for this survey. One would expect that there'd be an equal number of blue teams and purple teams with more honor badges, so any effect should cancel out. I mean, there's all sorts of variables I could try to account for if I wanted to be thorough, but my key assumption is that as the number of games gets very large, other factors should cancel out (like, number of 4v5 games, number of duo queues, etc).

Of course, my background is in physics, not statistics, so I could be wrong there.

My goal was to keep the submission form as simple as possible to make it easy for people to participate.
There should be an equal number of blue and purple teams with more badges; however, if SR is actually biased towards blue, the effect of honor might be over/underrepresented because of map asymmetry. This is especially true with large samples. Just food for thought. If honor does have significant effect you should still get good results!


Comment below rating threshold, click here to show it.

iatre

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

Senior Member

12-16-2012

Got my first two submissions (other than my own). Thanks you guys.


12