Earth: 2025 Talk

Earth: 2025 Talk / Alliance Talk / The Rank Project: The Final Ranking

Telstar
Member Jan 21st 12:06 PM
Well, I finished compiling the results of The Rank Project. This first post is nothing but an anti-break, so read on (or skip to the end) for the conclusion to this project. Oh, and cross your fingers and hope this doesn't break. :P

Telstar
Member Jan 21st 12:06 PM
THIS IS TOO LONG, WHY SHOULD I READ IT?

Don't, if you don't want to. Just check for the section titles and read what you care about. Check the end for the rankings without explanations if that's all you really care about.

HOW DID I COME UP WITH THIS LIST?

I took the initial list from the first The Rank Project thread, and took everyones suggestions and formed a new list from everyone's opinions.

Fortunately, the list wasn't ambiguous enough to where I needed to make personal decisions between "tied" alliances to determine who'd be on the list or not.

WHAT NEXT?

Now that Parts I, II and III (this one) are done, what happens now is up to whatever feels like a good idea. If anyone has any suggestions as to what could be done further with the data aquired through The Rank Project, feel free to say something.

Telstar
Member Jan 21st 12:07 PM
WHAT WAS THE ORIGINAL LIST?

This is the list that I first came up with, for which people voted alliances on or off the list until it formed the new list which was the basis for this project. This is in alphabetical order:

CWG
Elitez
FA
Genesis
Horizon
ICN
IX
LaF
LCN
League
Legacy
MD
NM
Omega
RAGE
Ragnarok
SoF
SOL
TIE
UCN

WHAT ARE THE UNORDERED TOP 20 ALLIANCES?

This is the list that resulted from people voting alliances on or off the list previously mentioned. For about a month, each day people were voting for what alliances they believed to be ranked below the alliance of that particular day. For example, on RAGE's day, people voted who they thought belonged below RAGE on a rankings list. Again, this is in alphabetical order:

AoD
Arrow
CWG
Elitez
Genesis
ICN
IX
LaF
League
Legacy
MD
NM
Omega
PoW
RAGE
SoF
SOL
TheCult
TIE
UCN

Telstar
Member Jan 21st 12:08 PM
HOW DID YOU DETERMINE THE FINAL RANKINGS?

First off, let me define some terminology so this doesn't end up being too wordy. The alliance that all other alliances are being compared to and appears in the topic of the Rank Project thread for a particular day, we'll call the "topic alliance". The other alliances being ranked compared to the topic alliance will be called the "compared alliances". A "vote" will signify each compared alliance individually. So each voter posts a list of several votes (the list of compared alliances). But no one person can get the same compared alliance counted more than once. In other words, a voter can post several alliances in their list, but not list the same alliance more than once (or post two lists).

So, I made a spreadsheet with topic alliances down the side and compared alliances across the top. For each topic alliance, I entered the number of votes each compared alliance received. In general, the more votes a compared alliance got, the better for the topic alliance, but worse for the compared alliance. So an alliance wants a high vote count across the spreadsheet (a high "topic alliance score" meaning a lot of people rank other alliances below you), but not down the spreadsheet (a high "compared alliance score" meaning a lot of people rank YOU below other alliances).

I then pumped out some statistics (mentioned in the following section) and made three lists from them.

Telstar
Member Jan 21st 12:08 PM
WHAT TYPES OF STATISTICS DID YOU COME UP WITH?

I'm glad you asked that. :) Here's three types of statistics that came out of this:

*(A)* The percent of votes an alliance received as a compared alliance out of the total number of people who voted for the corresponding topic alliance. For example, on IX's day, 9 people included NM in their list of alliances to be ranked below IX. There was a total of 11 people voting that day. That means 81.82% of the people that day voted for NM to be below IX. As a side note, on NM's day, 2 of 11 people voted for IX to be ranked below NM (18.18%).

*(B)* The sum of statistic (A) over all the days in the project, for a particular compared alliance. For example, on Arrow's day, 90.91% of the lists contained AoD. On LaF's day, 100% of the lists contained AoD. On NM's day, 90.91% of the lists contained AoD. Add those together, with all the other days, too, and you get 1365%. Divide that by 100 and you get 13.65. The way I look at that is since I can't just say AoD is ranked below Arrow and MD (I can say it about LaF) with absolute certainty, I can say instead that AoD is 90.91% ranked below Arrow, 100% ranked below LaF, and 90.91% ranked below NM. So rather than being ranked below 3 alliances, its ranked below 2.82 alliances (90.91 + 100 + 90.91 = 282).

*(C)* The total number of votes a compared alliance got (across the entire project) divided by the total number of votes it could have gotten (not counting the votes on the day it was a topic alliance, since you don't vote for an alliance to be ranked below itself). For example, a total of 163 lists of compared alliances (votes) were made across the entire project (some people voted more than once during the project, but not on the same day for the same topic alliance). But on RAGE's day, 9 lists were made, so RAGE could have, at most, only appeared on 154 lists. RAGE actually only appeared on 89 of those lists. That means RAGE appeared on 57.79% of the total number of lists it could have appeared on.

I formed three lists, each one based on a statistic. While the lists for the last two statistics are identical except for one has Arrow then PoW, and the other has PoW then Arrow, I put the most faith into the list formed by the first statistic. The way I formed the first list is by comparing each pair of alliances individually. Basically, if I look at NM and IX, and see that 81.82% of the people put IX ahead of NM, and only 18.18% of the people put NM ahead of IX, I'm going to put IX ahead of NM. Then I take another alliance and compare it with, say, NM and see if it belongs ahead or behind NM. If it belongs ahead of NM, I'll compare it with IX to see if it goes first, or between IX and NM. Continue this, and a list forms based on all these comparisons.

I went back and rechecked and found that with that list, if I check each alliance on there against each alliance listed ahead of it, the percent of people ranking it below the other alliance is always bigger than the people ranking the other alliance below it. If you can make sense of what I just said, that just means there are no contradictions of alliance A being ahead of alliance B, and alliance B being ahead of alliance C, but people thinking alliance C should be ahead of alliance A. The reason I put more faith into this one is because this one is based on direct comparisons of people's votes between two specific alliances, rather than an aggregate of data about how many votes an alliance received overall without consideration for just which alliance everyone was being ranked below.

Telstar
Member Jan 21st 12:10 PM
ENOUGH TALKING, WHAT ARE THE RESULTS?

Here's the list that I trust the most, based on statistic (A) mentioned in the previous section:

1 . UCN
2 . Elites
3 . SoF
4 . IX
5 . LaF
6 . MD
7 . Genesis
8 . SOL
9 . PoW
10. Legacy
11. NM
12. RAGE
13. Omega
14. Arrow
15. TIE
16. TheCult
17. League
18. AoD
19. ICN
20. CWG

To find out why I trust that list the most, read the previous section. :)

Telstar
Member Jan 21st 12:10 PM
Now, here's the list based on statistic (B):

1 . UCN ...... 2.02
2 . Elitez ... 2.28
3 . SoF ...... 3.33
4 . MD ....... 3.65
5 . LaF ...... 4.20
6 . IX ....... 4.72
7 . Genesis .. 5.04
8 . SOL ...... 7.79
9 . Legacy ... 8.35
10. RAGE ..... 8.55
11. NM ....... 8.86
12. Omega .... 9.25
13. TIE ...... 9.79
14. Arrow .... 9.95
15. PoW ...... 9.99
16. League .. 12.92
17. ICN ..... 13.50
18. CWG ..... 13.64
19. AoD ..... 13.65
20. TheCult . 13.82

The only major differences between this list and the last one are the positions of PoW and TheCult. Both are a few ranks lower in the second list. As for the rest, everyone's still in the same general area. Instead of IX, LaF, MD, for example, it's MD, LaF, IX. And instead of NM then RAGE, it's RAGE then NM. There's a couple other minor swaps like this, but overall it's not a huge difference.

Telstar
Member Jan 21st 12:11 PM
Here's the list based on statistic (C):

1 . UCN ..... 11.72%
2 . Elitez .. 14.57%
3 . SoF ..... 20.86%
4 . MD ...... 27.56%
5 . LaF ..... 30.00%
6 . IX ...... 32.24%
7 . Genesis . 33.74%
8 . SOL ..... 52.60%
9 . Legacy .. 56.21%
10. RAGE .... 57.79%
11. NM ...... 59.21%
12. Omega ... 61.69%
13. TIE ..... 63.46%
14. PoW ..... 64.33%
15. Arrow ... 65.13%
16. League .. 81.82%
17. ICN ..... 83.75%
18. CWG ..... 84.91%
19. AoD ..... 86.16%
20. TheCult . 86.93%

As you can see, the ONLY difference between this one and the last one is the position of Arrow and PoW. However, these last two lists focus more on how many votes the alliances have received overall rather than a specific comparison between every alliance pair to find the correct order. The first list *is* based on a comparison between each alliance pair, and there weren't any contradictions, however I should note that there were two "tied" alliance pairs where the same percentage of people said one should be below the other as the other way around. However, those ties were resolved during the course of comparing with other alliances on the list.

Telstar
Member Jan 21st 12:12 PM
WHAT ABOUT AoD? THEY...

I know, I know. ;) Pretend they still exist.

ARE YOU GOING TO SAY SOMETHING ABOUT THE ALLIANCES? ANY OPINIONS?

*I* don't have any opinions. However, I do have the opinions of those who cared to write something up about their own alliance.

Telstar
Member Jan 21st 12:12 PM

This is a simulated thread... I didn't actually save the original thread from AT, but I did save the text I posted on AT that day, so you won't see any responses from people here. As for the opinions on alliances, see the Thoughts on Alliances section from the main index.

Telstar Member	Jan 21st 12:06 PM Well, I finished compiling the results of The Rank Project. This first post is nothing but an anti-break, so read on (or skip to the end) for the conclusion to this project. Oh, and cross your fingers and hope this doesn't break. :P
Telstar Member	Jan 21st 12:06 PM THIS IS TOO LONG, WHY SHOULD I READ IT? Don't, if you don't want to. Just check for the section titles and read what you care about. Check the end for the rankings without explanations if that's all you really care about. HOW DID I COME UP WITH THIS LIST? I took the initial list from the first The Rank Project thread, and took everyones suggestions and formed a new list from everyone's opinions. Fortunately, the list wasn't ambiguous enough to where I needed to make personal decisions between "tied" alliances to determine who'd be on the list or not. WHAT NEXT? Now that Parts I, II and III (this one) are done, what happens now is up to whatever feels like a good idea. If anyone has any suggestions as to what could be done further with the data aquired through The Rank Project, feel free to say something.
Telstar Member	Jan 21st 12:07 PM WHAT WAS THE ORIGINAL LIST? This is the list that I first came up with, for which people voted alliances on or off the list until it formed the new list which was the basis for this project. This is in alphabetical order: CWG Elitez FA Genesis Horizon ICN IX LaF LCN League Legacy MD NM Omega RAGE Ragnarok SoF SOL TIE UCN WHAT ARE THE UNORDERED TOP 20 ALLIANCES? This is the list that resulted from people voting alliances on or off the list previously mentioned. For about a month, each day people were voting for what alliances they believed to be ranked below the alliance of that particular day. For example, on RAGE's day, people voted who they thought belonged below RAGE on a rankings list. Again, this is in alphabetical order: AoD Arrow CWG Elitez Genesis ICN IX LaF League Legacy MD NM Omega PoW RAGE SoF SOL TheCult TIE UCN
Telstar Member	Jan 21st 12:08 PM HOW DID YOU DETERMINE THE FINAL RANKINGS? First off, let me define some terminology so this doesn't end up being too wordy. The alliance that all other alliances are being compared to and appears in the topic of the Rank Project thread for a particular day, we'll call the "topic alliance". The other alliances being ranked compared to the topic alliance will be called the "compared alliances". A "vote" will signify each compared alliance individually. So each voter posts a list of several votes (the list of compared alliances). But no one person can get the same compared alliance counted more than once. In other words, a voter can post several alliances in their list, but not list the same alliance more than once (or post two lists). So, I made a spreadsheet with topic alliances down the side and compared alliances across the top. For each topic alliance, I entered the number of votes each compared alliance received. In general, the more votes a compared alliance got, the better for the topic alliance, but worse for the compared alliance. So an alliance wants a high vote count across the spreadsheet (a high "topic alliance score" meaning a lot of people rank other alliances below you), but not down the spreadsheet (a high "compared alliance score" meaning a lot of people rank YOU below other alliances). I then pumped out some statistics (mentioned in the following section) and made three lists from them.
Telstar Member	Jan 21st 12:08 PM WHAT TYPES OF STATISTICS DID YOU COME UP WITH? I'm glad you asked that. :) Here's three types of statistics that came out of this: (A) The percent of votes an alliance received as a compared alliance out of the total number of people who voted for the corresponding topic alliance. For example, on IX's day, 9 people included NM in their list of alliances to be ranked below IX. There was a total of 11 people voting that day. That means 81.82% of the people that day voted for NM to be below IX. As a side note, on NM's day, 2 of 11 people voted for IX to be ranked below NM (18.18%). (B) The sum of statistic (A) over all the days in the project, for a particular compared alliance. For example, on Arrow's day, 90.91% of the lists contained AoD. On LaF's day, 100% of the lists contained AoD. On NM's day, 90.91% of the lists contained AoD. Add those together, with all the other days, too, and you get 1365%. Divide that by 100 and you get 13.65. The way I look at that is since I can't just say AoD is ranked below Arrow and MD (I can say it about LaF) with absolute certainty, I can say instead that AoD is 90.91% ranked below Arrow, 100% ranked below LaF, and 90.91% ranked below NM. So rather than being ranked below 3 alliances, its ranked below 2.82 alliances (90.91 + 100 + 90.91 = 282). (C) The total number of votes a compared alliance got (across the entire project) divided by the total number of votes it could have gotten (not counting the votes on the day it was a topic alliance, since you don't vote for an alliance to be ranked below itself). For example, a total of 163 lists of compared alliances (votes) were made across the entire project (some people voted more than once during the project, but not on the same day for the same topic alliance). But on RAGE's day, 9 lists were made, so RAGE could have, at most, only appeared on 154 lists. RAGE actually only appeared on 89 of those lists. That means RAGE appeared on 57.79% of the total number of lists it could have appeared on. I formed three lists, each one based on a statistic. While the lists for the last two statistics are identical except for one has Arrow then PoW, and the other has PoW then Arrow, I put the most faith into the list formed by the first statistic. The way I formed the first list is by comparing each pair of alliances individually. Basically, if I look at NM and IX, and see that 81.82% of the people put IX ahead of NM, and only 18.18% of the people put NM ahead of IX, I'm going to put IX ahead of NM. Then I take another alliance and compare it with, say, NM and see if it belongs ahead or behind NM. If it belongs ahead of NM, I'll compare it with IX to see if it goes first, or between IX and NM. Continue this, and a list forms based on all these comparisons. I went back and rechecked and found that with that list, if I check each alliance on there against each alliance listed ahead of it, the percent of people ranking it below the other alliance is always bigger than the people ranking the other alliance below it. If you can make sense of what I just said, that just means there are no contradictions of alliance A being ahead of alliance B, and alliance B being ahead of alliance C, but people thinking alliance C should be ahead of alliance A. The reason I put more faith into this one is because this one is based on direct comparisons of people's votes between two specific alliances, rather than an aggregate of data about how many votes an alliance received overall without consideration for just which alliance everyone was being ranked below.
Telstar Member	Jan 21st 12:10 PM ENOUGH TALKING, WHAT ARE THE RESULTS? Here's the list that I trust the most, based on statistic (A) mentioned in the previous section: 1 . UCN 2 . Elites 3 . SoF 4 . IX 5 . LaF 6 . MD 7 . Genesis 8 . SOL 9 . PoW 10. Legacy 11. NM 12. RAGE 13. Omega 14. Arrow 15. TIE 16. TheCult 17. League 18. AoD 19. ICN 20. CWG To find out why I trust that list the most, read the previous section. :)
Telstar Member	Jan 21st 12:10 PM Now, here's the list based on statistic (B): 1 . UCN ...... 2.02 2 . Elitez ... 2.28 3 . SoF ...... 3.33 4 . MD ....... 3.65 5 . LaF ...... 4.20 6 . IX ....... 4.72 7 . Genesis .. 5.04 8 . SOL ...... 7.79 9 . Legacy ... 8.35 10. RAGE ..... 8.55 11. NM ....... 8.86 12. Omega .... 9.25 13. TIE ...... 9.79 14. Arrow .... 9.95 15. PoW ...... 9.99 16. League .. 12.92 17. ICN ..... 13.50 18. CWG ..... 13.64 19. AoD ..... 13.65 20. TheCult . 13.82 The only major differences between this list and the last one are the positions of PoW and TheCult. Both are a few ranks lower in the second list. As for the rest, everyone's still in the same general area. Instead of IX, LaF, MD, for example, it's MD, LaF, IX. And instead of NM then RAGE, it's RAGE then NM. There's a couple other minor swaps like this, but overall it's not a huge difference.
Telstar Member	Jan 21st 12:11 PM Here's the list based on statistic (C): 1 . UCN ..... 11.72% 2 . Elitez .. 14.57% 3 . SoF ..... 20.86% 4 . MD ...... 27.56% 5 . LaF ..... 30.00% 6 . IX ...... 32.24% 7 . Genesis . 33.74% 8 . SOL ..... 52.60% 9 . Legacy .. 56.21% 10. RAGE .... 57.79% 11. NM ...... 59.21% 12. Omega ... 61.69% 13. TIE ..... 63.46% 14. PoW ..... 64.33% 15. Arrow ... 65.13% 16. League .. 81.82% 17. ICN ..... 83.75% 18. CWG ..... 84.91% 19. AoD ..... 86.16% 20. TheCult . 86.93% As you can see, the ONLY difference between this one and the last one is the position of Arrow and PoW. However, these last two lists focus more on how many votes the alliances have received overall rather than a specific comparison between every alliance pair to find the correct order. The first list is based on a comparison between each alliance pair, and there weren't any contradictions, however I should note that there were two "tied" alliance pairs where the same percentage of people said one should be below the other as the other way around. However, those ties were resolved during the course of comparing with other alliances on the list.
Telstar Member	Jan 21st 12:12 PM WHAT ABOUT AoD? THEY... I know, I know. ;) Pretend they still exist. ARE YOU GOING TO SAY SOMETHING ABOUT THE ALLIANCES? ANY OPINIONS? I don't have any opinions. However, I do have the opinions of those who cared to write something up about their own alliance.
Telstar Member	Jan 21st 12:12 PM This is a simulated thread... I didn't actually save the original thread from AT, but I did save the text I posted on AT that day, so you won't see any responses from people here. As for the opinions on alliances, see the Thoughts on Alliances section from the main index.