Jump to content
PSK

Mafia Rankings

Recommended Posts

Jazz    5331
1 minute ago, ZeroPassion said:

 

this is explained in frank's post of the elo system above - only based on an actual two team game of town vs mafia as elo is based on a 1v1. generally, this is done as a person v person but it should work out through a long series of games for each individual on the team for their individual elo instead. 

 

I understand that, I'm just pointing out the irony of the circumstances.

Share this post


Link to post
Share on other sites
Solstice    390
1 hour ago, ZeroPassion said:

I've decided to take the opportunity to continue the ELO system based on the rankings thread and start from games in January '17 that do not include Neutrals.

 

So far I'm through 3 games with just the raw numbers excluding any MVP, Modkill as of now since that had not been figured out yet.

 

https://docs.google.com/spreadsheets/d/1X7FVPiKbN6PDxhJfhel7eNNsTXyRI5gfOTnn1up6db4/edit?usp=sharing 

 

Elo is on the first sheet, the numbers come from the second sheet, and the notes on how I'm doing the whole thing are in the third sheet.

 

Already @Jazz is the shittiest player on the forums through the first 3 recorded games! Congratulations on 3 games proving your worth! 

 

But these numbers really are a pain in the ass. Someone should help in collaborating with me. 

 

I'm ranked higher than Malcolm!?

 

...The apocalypse is surely coming.  

Edited by Solstice

Share this post


Link to post
Share on other sites
ZeroPassion    706
18 minutes ago, Solstice said:

 

I'm ranked higher than Malcolm!?

 

...The apocalypse is surely coming.  

 

i may or may not disappoint you when i finish this tomorrow! 

Share this post


Link to post
Share on other sites
+scuzzlebutt    23495
7 hours ago, my bloody arcanine said:

it's been over a year since i last played and there isn't a decent ranking system yet?

eLo IsNt ObJeCtIvE aNd DoEsNt WoRk :^]]]

  • Downvote 1

Share this post


Link to post
Share on other sites
+scuzzlebutt    23495

btw hey even tho i made fun of ur opinion try not to have a conniption fit and unfriend a bunch of people on facebook ok???

  • Downvote 1

Share this post


Link to post
Share on other sites
»JC.    4973

is there really anything to gain from objective rankings besides letting people go "haha ur lower than me u suk"

Share this post


Link to post
Share on other sites
Jazz    5331

ELO for mafia is pretty dumb from a statistical perspective. First you're trying to use small sample sizes (a typical player plays 15 or fewer games per year) to detect small differences between players who have similar base skill levels. Then you have to consider all the noise that's caused by random factors such as hosting, teammates' play, bad balance, neutrals, quirks, etc.

 

We don't really care about using the ranking system to tell us that Malcolm is better than Tyranno. That's all an ELO system will be able to tell you, hence why nobody who actually plays and takes this game seriously gives a shit.

 

Even if you make the assumption that the random factors I listed are non-existent or small (which is a bad assumption), I highly doubt you get a statistically significant result between players who are actually close in ability, which is what your goal ought to be if you are advocating a statistics based approach.

 

Like JC's thread is cool and all but I think slick's arbitrary rankings are far more accurate and far more fun to discuss.

  • Upvote 1

Share this post


Link to post
Share on other sites
»JC.    4973
1 hour ago, Jazz said:

We don't really care about using the ranking system to tell us that Malcolm is better than Tyranno

 

Like JC's thread is cool and all but I think slick's arbitrary rankings are far more accurate and far more fun to discuss.

 

C Tier (<50%)
TheGoldenTyranno (43%)
Wunterslaus (42%)

D Tier (< 40%)
Malcolm (31%)

 

my stat based rankings have no flaws stfu

 

  • Upvote 2

Share this post


Link to post
Share on other sites
+scuzzlebutt    23495
18 hours ago, Jazz said:

ELO for mafia is pretty dumb from a statistical perspective. First you're trying to use small sample sizes (a typical player plays 15 or fewer games per year) to detect small differences between players who have similar base skill levels. Then you have to consider all the noise that's caused by random factors such as hosting, teammates' play, bad balance, neutrals, quirks, etc.

 

We don't really care about using the ranking system to tell us that Malcolm is better than Tyranno. That's all an ELO system will be able to tell you, hence why nobody who actually plays and takes this game seriously gives a shit.

 

Even if you make the assumption that the random factors I listed are non-existent or small (which is a bad assumption), I highly doubt you get a statistically significant result between players who are actually close in ability, which is what your goal ought to be if you are advocating a statistics based approach.

 

Like JC's thread is cool and all but I think slick's arbitrary rankings are far more accurate and far more fun to discuss.

1. most of you guys have played 15 games in the first half of this year

 

2. "players who have similar base skill levels" [citation needed]

 

3. whose fault is it that u guys are more concerned with designing a 5-faction 70-power role shitfest than actually playing something that might test skills that are relevant to the game of mafia

 

4. malcolm < tyranno fite me irl

 

5. "look at all the RANDOM FACTORS, how are u ever gonna rank skill in a game like this???? mafia is so unique guise :^)))"

 

http://www.globalpokerindex.com/

https://champtrade.com/games/yugioh#/ChampRank?_k=cuwocd

http://duelistgroundz.com/index.php?/topic/170205-the-ratingsrankings-thread/&page=1

 

6. slicks' thread is a power ranking, not a true rating system. both have pros and cons (and ur not fooling me, i know u just like them because they have u above mark!)

 

like why cant u guys just be honest and say you dont want to do elo or anything other than semi-serious arbitrary PRs bc it's too much work (that ZP and JC are already willing to do for u anyways...)

 

18 hours ago, my bloody arcanine said:

who is this random anyway?

 

trump-laughing-1.jpg

 

 

  • Upvote 1

Share this post


Link to post
Share on other sites
Jazz    5331

I think the stats will only be good for super active players. Maybe my rating would be accurate because I sign up most of the time.

 

We have had exactly 25 completed games so far this year (according to JCs thread).

 

I have hosted 4. I was banned for 4. I missed 2 others because they were double games. So I have played in 15 games and I'm one of the most active players on the site.

 

I don't think this is a big enough sample size.

 

BTW we introduce a huge statistical problem in unequal sample sizes between players. You'll have some players with 5 games played and others with 15+ which is not good for comparisons.

Share this post


Link to post
Share on other sites
Jazz    5331

Citation for players close in win percentage is the assumption that skill is normally distributed and normal distributions have a cluster of scores in the middle. Boom.

 

(also the win rates as calculated by JC)

Share this post


Link to post
Share on other sites
+scuzzlebutt    23495

25 sounds fine to me if everyone is ok with including weird boundary-pushing games like GoT (which didnt seem to be a problem for jc). if the year finishes out at the same rate you'd end up with more than 30. consider how relatively small the section is; you are extremely active and play in most games whenever possible and many other players are as well. there's also no rule that says you have to limit it to just this past year (as long as people are willing to dig)

Share this post


Link to post
Share on other sites
+Sophocles    3079
17 minutes ago, mmf said:

 

3. whose fault is it that u guys are more concerned with designing a 5-faction 70-power role shitfest than actually playing something that might test skills that are relevant to the game of mafia

 

 

 

 

have fun -or- spend a lot of time designing a rating system that probably is heavily flawed anyway

 

hmm...

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


  • Recently Browsing   0 members

    No registered users viewing this page.

×