Jump to content

What PnW Alliance Are You?: A Personality Quiz and WoT


hidude45454
 Share

Recommended Posts

UPN is right next to oblivion and test in the chart? Proof that UPN is an upper-tier hegemon 

  • Haha 2

Orbis Wars   |   CSI: UPN   |   B I G O O F   |   PW Expert Has Nerve To Tell You How To Run Your Own Goddamn Alliance | Occupy Wall Street | Sheepy Sings

TheNG - My favorite part is when Steve suggests DEIC might have done something remotely successful, then gets massively shit on for proposing such a stupid idea.

On 1/4/2016 at 6:37 PM, Sheepy said:
Sheepy said:

I'm retarded, you win

Link to comment
Share on other sites

1. Carthago: 80% similarity {RIP}

2. The Hand of Fate: 72% similarity

3. Black Knights: 72% similarity

4. Cataclysm: 70% similarity

5. House Stark: 69% similarity

Bottom:

1. Yarr: 37% similarity

2. Respublica Romana: 39% similarity

3. Farkistan: 42% similarity

 

Hmm.. Somewhat accurate, but tbqh I picked a lot of neutral options, you gotta add more 'personality' options in your quiz to really narrow the options down imho.

  • Upvote 1
Link to comment
Share on other sites

Top 5

1. Carthago: 77% similarity

2. The Knights Radiant: 72% similarity

3. Arrgh: 69% similarity

4. Eclipse: 67% similarity

5. The Hand of Fate: 67% similarity

Bottom 5

1. Yarr: 23% similarity

2. Respublica Romana: 27% similarity

3. Genesis: 29% similarity

4. Farkistan: 30% similarity

5. Sanctuary: 35% similarity

I'm both surprised and not surprised by my results

Link to comment
Share on other sites

1. CTO
2. Weeb
3. Eclipse(wot)
4. TKR(been there done that)
5. Aurora

Bottom 5
1.Yarr
2. RR
3. Fark
4.Morningstar
5. Genesis 

               

Link to comment
Share on other sites

Thanks for the positive reception y'all, huge relief a lot of people are actually getting some reasonable results out of this ❤️

I have edited the original post with a changelog so people can see how I'm updating the project! As of now, I fixed some alliance descriptions, obfuscated some code, and most importantly added two possible ways to embed your results for easy sharing so you don't have to screenshot the entire page. As a demonstration, here are my own personal results:

Quote

Alliance Personality Test by hidude45454

My Top 3: 🥇 Oblivion (77%) / 🥈 HoF (76%) / 🥉 TEst (75%)
My Bottom 3: 🥇 Fark (43%) / 🥈 Genesis (45%) / 🥉 Yarr (45%)

image.png.0f201681158584c5f864e1378d387982.png

Yes, I've noticed a lot of people have CTO somewhere in their top 5! My best guess as to why is CTO may just be very middle of the road for what most people think most alliances "should" be, even more so for the people that talk in news servers and public chats that may share some common biases.

Link to comment
Share on other sites

On 5/21/2022 at 9:07 PM, hidude45454 said:

FOR PEOPLE WHO JUST WANT THE QUIZ AND DON'T CARE ABOUT THE ANALYSIS:

https://alliance-personality-quiz.herokuapp.com/

The TL;DR:

13 questions to give you an estimate of your top 5 and bottom 5 most/least similar alliances, maybe useful for players looking to join an alliance! 4 of those questions are based on my rough assessment of each alliance, and 1 is based on low-tier raiding rates (will cover both in detail below), so feel free to exclude those questions if you want. This is only semi-scientific so please treat as primarily entertainment :)

Alliances Counted:

All the alliances that I contacted were in the top 50 or training alliances of alliances in the top 50 (I separated training aas from main aas).

Alliances that got back to me on my questionnaire (thank you to the respondents!):

Rose, The Knights Radiant, The Syndicate, The Immortals, Cataclysm, The Fighting Pacifists, Eclipse, The Commonwealth, Bourbon Street, Grumpy Old Bastards, Guardian, Aurora, Carthago, Terminus Est, Weebunism, The Hand of Fate, The Legion, The Wei, Black Knights, Soldiers of Liberty, Dark Brotherhood, House Stark, Waffle House, Camelot, Name Withheld, Order of the White Rose, The Dead Rabbits, United Purple Nations, The Enterprise, The Elites, Federated States of Orbis, Serpentis, The Armed Peacekeepers, Arrgh, Farkistan, Church Of Atom, Global Alliance & Treaty Organization, Polaris, Hogwarts, Unforgiven Legion, United Socialist Nations, Yarr, The Federation, Union of Soviet Socialist Republics, Respublica Romana, Genesis, Sanctuary, Morningstar, Oblivion, Apollo, Otaku Shougaku, The Mortals, Convent of Atom, The Circus, The Cosmonauts, Daedalus, The Hive

Pending alliances:

None at the moment!

Declined alliances:

Advanced Syndicalist Mechanics

This is a running quiz, meaning I will constantly be adding/modifying alliances and alliance results, so if you would like your alliance featured or would like your alliance values changed please send me a message or DM (I will include any alliances regardless of rank!) If you would like your final alliance description to be changed, please contact me as well! Alliance responses will be kept confidential.

Methodology:

13 questions were chosen based on general traits and descriptions you may recognize each specific alliance for, including around areas of basic info, structure, community, and competence. I will provide a description of each question/trait below! 7 of those questions were picked based on factual information from each alliance, and the other 6 were directly taken from questions sent to an IA gov member, close contact, or leader in each alliance:

The structure of the quiz was directly inspired from The Official Knights Radiant Order Quiz (AKA: the thing you take to get into TKR) and tl;dr no quiz site on the market had a way to set up a quiz like this on their own site so I legit just coded it and made my own site rofl. (I hope it looks okay)

Potential downsides: some people may prioritize certain traits more than others, and that wasn't accounted for, but that would probably take a lot of additional effort to factor in. I will do it if people request it enough though. Some people also may prefer a Likert scale over a 100-point scale, but too bad :serious:

Question/Trait 1: Size

How it was measured: Each alliance was given a score averaged from its score ranking and its membership count ranking to give equal weight to both (this was done pre-militarization), and this score was compared to all other values.

Potential downsides: This is just a heuristic and not any official measurement. Since main aas/training aas/extensions were counted separately, there may be some fuzziness with membership/score counts possibly combining between them.

Question/Trait 2: Age

How it was measured: The age of the alliance was calculated and compared to the age of other alliances.

Potential downsides: Different people may perceive what "old" and "new" are differently depending on stuff like how long they've been playing the game.

Question/Trait 3: Competency [HIDUDE-BIASED]

How it was measured: Each alliance was given a grade based on my assessment of their general war performance, and this grade was converted into a score.

Potential downsides: My assessment is obviously biased in this aspect (although I think it's pretty good), and also the grade can't perfectly be converted into a score.

Question/Trait 4: Profile [HIDUDE-BIASED]

How it was measured: Each alliance was given an estimated value based on my assessment of how public or visible the alliance tries to be and/or is made so by the community.

Potential downsides: Same as above, my perception of profile will be different than other people's.

Question/Trait 5: Exclusitivity, Question/Trait 6: Maturity, Question/Trait 7: Technology, Question/Trait 8: RP, Question/Trait 9: Activity, Question/Trait 10: Community

How it was measured: These were directly copied from the responses alliance representatives gave me.

Potential downsides: Due to me keeping responses anonymous, alliances may not have had a perfect baseline to compare their responses to, and also may have felt the urge to inflate certain values to make themselves look more desirable, although I can't know to what degree. Still, I think people did the best job they could answering these.

Question/Trait 11: War Rate [HIDUDE-BIASED]

How it was measured: I counted the number of offensive wars each alliance has declared in the past two years and translated that into a rough assessment of how they prefer to fight.

Potential downsides: Same as above, my perception of war will be different than other people's.

Question/Trait 12: FA Leadership [HIDUDE-BIASED]

How it was measured: I gave a value based on my assessment of each alliances' major/macro/micro status and how much they participated in FA discussions.

Potential downsides: Same as above, my perception of FA presence will be different than other people's, especially since I obviously don't have a full picture here.

Question/Trait 13: Raid Rate [LOW-TIER RAIDING]

How it was measured: I wrote a script (pre-war) to measure every alliance's number of wars C10 and below for the past month, and counted the number of members C10 and under to get a number of wars declared per person:

unknown.png

This rate was compared to other rates and given a value.

Potential downsides: As you can see, a lot of big alliances have 0s, primarily because they don't have any members C10 or below. This is why I made the low-tier raiding question optional in my quiz.

Percent Similarity Calculations:

For all % similarities I displayed, I calculated the Euclidean distance between a user's responses and the responses of each alliance, and divided by the max possible Euclidean distance to get a "dissimilarity" score, which I could convert into a similarity percentage. (Not the most scientific way of doing it, but there are a bunch of different methods of doing it too that I considered not applicable, although feel free to suggest another way of doing it)

Alliance Comparison:

I decided to take my data one step further and compare alliances with each other to see the most similar and least alliances to each alliance. In order to do this, I also created separate categories of classification for comparison:

General: includes all alliance traits.

Alliance Structure: how an alliance's government is set up and how the alliance runs in general. For example, democracies may run differently than dictatorships, and older alliances may run different than newer ones. Includes traits #1, 2, 3, 5, 7, 8, and 13.

Impact: how much of an impact an alliance strives to have on the game as a whole. More impactful alliances generally seek to have a greater public image or FA sector and/or can punch above their weight during war. Includes traits #1, 3, 4, 9, 11, 12, and 13.

Community: what type of community each alliance has. Based on the responses I got from representatives. Includes traits #5, 6, 8, 9, and 10.

Competence: the military or political ability of each alliance. Includes traits #3, 4, 5, 9, and 11.

Seniority: the age and experience of each alliance. For example, more senior alliances are older in age (duh), may have more mature members, and focus more on non-PnW discussion, and vice versa for newer alliances. Includes traits #2, 6, and 10.

You can see how each alliance stacked up here: https://docs.google.com/spreadsheets/d/1FerD_fKtu-5kdxDGlTYRo23eAWmyM2VdS09mFsW-LaQ/edit#gid=822663440

Grouping Alliances:

Finally, I decided to group all alliances as a whole for fun. This was done through a process called dimensionality reduction so that I could reduce a 13d question space into a 2d space easily interpretable by humans.

First, I processed question data and performed PCA on it.

The tl;dr of what PCA does is it uses your original responses as "weights" in order to determine a combined set of important generalized features of each alliance. Here is a percentage graph of how significant each feature was in the final calculation:

image.png.55d2e4a329e590c1a1c3a72f32e101af.png

And here's how each question "weight" or factored into each of the first three features:

image.png.4a4430fab13ff4b30b5ef97c9985e131.png

Finally, in order to arrange all the alliances into identifiable groups, I picked an optimal number of groupings to use by processing the data using an elbow method for clustering:

image.png.188bf0eaa8e9b540d70c197651d9ee3c.png

Using that, I decided to create about 4 separate groups. Using the chart generated from PCA, I clustered into 4 groups using K-means clustering in order to get this final chart:

image.thumb.png.a4df7971ef6d43f973ee6df2cabbf505.png

Here, you can see some interesting results and kind of see how each alliance got grouped, even though some questions may not have covered those traits directly!

Light Blue: this seems to be a grouping of generally higher-tier or more exclusive older alliances.

Pink: this seems to be a grouping of mostly lower-tier or mid-tier alliances that I can only graciously describe as "mid" lmao. Most of them belong to Backrooms/Johnsons/HM.

Dark Blue: this seems to be a grouping of mostly lower-tier alliances that are either fairly new/low-profile or training alliances.

Red: this seems to be a grouping of mostly mid-tier alliances that often have long histories or who are fairly established and well-known in-game.

Final Thoughts

I certainly learned a lot from this entire process, both from the coding aspect and finding out more about each alliance! Thanks to all the people who read through the entire way, and I hope results are to people's liking :) Again, please contact me if you would like your alliance values added or changed!

Changelog

5/22/22:

-Fixed descriptions for WH, TI, TEst, and USN

-Obfuscated code

-Added shareable embed/image options for Discord or Slack

-Added Hive

Are you taking a data science class?  
 

I like the effort you put in to attempt this as an unsupervised learning problem.  I’d be a bit careful here though because your first two principal components only capture about 40% of variance from what I can see.  
 

Also, just for curiosity sake what was your distance metric for k-means? 

Link to comment
Share on other sites

59 minutes ago, Cooper_ said:

Are you taking a data science class?  
 

I like the effort you put in to attempt this as an unsupervised learning problem.  I’d be a bit careful here though because your first two principal components only capture about 40% of variance from what I can see.  
 

Also, just for curiosity sake what was your distance metric for k-means? 

Thanks! Took a data science class about a year ago but this is the first time I'm applying some things from it outside it (plus a few additional things I picked up).

Yeah, admittedly translating this directly to two dimensions misses a bit of information, but I think it's mostly okay for the purposes of entertainment. Though, as an example if you bump it up to 5 features or so which captures about 70% of the variance, then the clustering looks a bit more like this:

image.thumb.png.914f2c6f577745238d6c60424aaa973f.png

And for k-means it's using Euclidean distances.

Link to comment
Share on other sites

10 hours ago, hidude45454 said:

Thanks! Took a data science class about a year ago but this is the first time I'm applying some things from it outside it (plus a few additional things I picked up).

Yeah, admittedly translating this directly to two dimensions misses a bit of information, but I think it's mostly okay for the purposes of entertainment. Though, as an example if you bump it up to 5 features or so which captures about 70% of the variance, then the clustering looks a bit more like this:

image.thumb.png.914f2c6f577745238d6c60424aaa973f.png

And for k-means it's using Euclidean distances.

Cool man!  I’m studying bioinformatics and focus on machine learning.  If you ever want to chat or something, reach out! 
 

I really appreciate you taking a more rigorous approach to your analyses, and thanks for the 5 PC graph!

Link to comment
Share on other sites

15 hours ago, rDaRealSpeedy said:

Why was ASM declined?

They probably declined to be included in the quiz for whatever reason. Tho maybe Hidude secretly has an anti-ASM agenda.

Link to comment
Share on other sites

26 minutes ago, Mayor said:

They probably declined to be included in the quiz for whatever reason. Tho maybe Hidude secretly has an anti-ASM agenda.

It was the former, FA apparently decided to not participate for some reason.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use and the Guidelines of the game and community.