Introducing K = constant of competitiveness for CW per seasons

I dont know why Croat deleted their thread but i have a possible solution in an analytical approach

I won't repost the stats that we actually have on this season, but as many coalitions have high Elo even with many few games played, let me propose 2 constants or ways to empirically estimate which coalitions should qualify to playoffs taking in consideration all the variables we have. I did some experiments and i have come with 2 plausible new constants to measure the degree of competitiveness of a coalition on any time of any season,using this in then could make it more fair to select coalitions that should qualify to playoffs:

The constant i'm proposing to do this is:



in which k is the constant of competitiveness
w is the quantity of CW's won
n is the number of games played
and E is the current season Elo.

Doing the math for each coalition we have:



Therefore, we can make this graph:



And with this empyric expression is obvious which coalitions can qualify to the playoffs while making this constant a fair way to calculate it taking in consideration all the key constants in the game.

Comment and give feedback if you like this new constant or no.

I deleted it because I wanted to re-do it and let WD to provide his idea which I like the most (we just had some longer talk and summarized it all up and it's pretty promissing). It's basically so simple solution to these outcomes and will award actual competing enough so active clans (like PRES, ENIG, HELLE, BD) don't feel damaged for actual playing.
Will let him to post and explain it to you in better way, as it's his own idea which is the one I support.
It could be implemented for future seasons (too late for current one).


Also, I liked the way how you found a correlation between all competitive aspects and made your formula , but what I don't like is deciding anything through formula, especially not using it to set up finalists. Even it includes aspects such as number of wins and Elo, it still makes leaderboard basically useless. This season finishes soon, it's too late (and unfair) to set up only the amount of CWs needed to qualify Finals now, so we could just extend the Final4 with clans who are positioned 1-4, plus adding clans who played more than lets say 20 cws into Final tournament.
Those clans who are not competing at all will drop off by themselves in the Final tournament and the outcome will be similar.
It's the simplest solution for this season, imo and no one is hurted or restricted.

I like the general idea, but the equation should be modified. We should keep the same terms in the numerator/denominator (so we're not changing much), but some of the operations should change:



Such that as you approach 1, your qualification increases

My proposal is to give reward participation by awarding bonus elo. If you look at the leaderboard right now, Paradox has a huge chunk of elo and you have clans such as hellenism and enigma who are sub 1000 elo but are very active. This makes sense because in order for a clan to climb above 1000 elo, another clan has to drop below 1000.

Add in the fact that we dont have new clans joining the cw season to increase the elo pool and we have the current problem we have now.

Basically the way this works is, in every clan war, regardless of win/loss, you get +1 added to your elo gain or elo loss. This will gradually increase the elo pool over the season and make it possible for active clans to climb out of the 900 elo hellhole an catch up to the #1 clan.

Here's a simple scenario to illustrate this. Assume there are 3 clans: A,B,C and that clan A cwed clan B 50 times and have exactly 50/50 win rate. Then at the last day of the season all clans have 1000 elo. Then clan B attempts to cw clan C but loses.

This would be the standings using the current system.

Clan C: 1st 1015 elo 1 game
Clan A: 2nd 1000 elo 50 games
Clan B: 3rd 985 elo 51 games.

This is a result of the fact that the elo must add up to 3000. So far some clans to rise above 1000, another clan has to drop below. This somewhat reflects the current leaderboard standing. Paradox owns so much of the elo pool that it is practically impossible for enigma, prestige, hellenism to climb out of 900 elo hell. At most only 2 of them can climb out of elo hell by farming the third clan into 800 elo. What most likely will happen is only one of the clan will climb out of 900 and the other two had to feed them to get out.

If we tried the new system with the same scenario, the end of the season clan A and B would have 1050 elo each. Clan B once again tries to cw clan C but loses. This would be the final standing.

Clan A: 1st 1050 elo 50 games
Clan B: 2nd 1036 elo 51 games
Clan C: 3rd 1016 elo 1 game.

With the current system, clan C cannot win because they won a single match. If we look at the current leaderboard, there would easily be an extra 300 elo in the elo pool, divided up among the active clans. This means that clans with one game like Backup cant win the season over a clan that's a lot more active. Unlike duels where in theory you have a fresh supply of elo from new players, clans take a lot longer to form and getting new clans to cw is a lot harder. So, the elo pool remains stagnant and we have the current leaderboard issues. With this elo bonus it will be possible for clans to be above 1000 elo without forcing another clan to drop below to 900 elo hell.

लिखा द्वारा Witch-Doctor, 04.02.2022 at 14:09

So is the 1 elo bonus given out after each cw or at the end of the season?

लिखा द्वारा Tungston., 04.02.2022 at 15:29

So is the 1 elo bonus given out after each cw or at the end of the season?

After each cw obviously and it's simple solution to avoid these outcomes for next seasons.
For current season, we can go on something like making bigger Final tournament, for example: Top 4 positioned clans + clans who played more than 20 games.

This is an interesting idea, but I don't think the math would work as described. The elo bonus won't be the seperate thing that sits there and accumilates, it will just be dumped into the current elo system and constantly redistrubuted. While increasing the elo pool would increase everyones elo some, it wouldn't increase it enough to fix the standings problems. I ran some numbers. Lets say we start juicing the current cw season with this elo, an extra 316 elo (games played this season) isn't going to push any of the current active clans mention above 1000 elo again, 316 elo isn't enough to do that i don't think. Can someone explain how the math would accomplish this, cause saying they played 50 cws and adding 50 elo isn't how it would work. ELO gets redistrubuted and wouldn't a lot of the extra elo just end up at the top with paradox and not really fix these standing problems? IDK, someone smarter than me figure this out.

लिखा द्वारा Tungston., 05.02.2022 at 08:50

but I don't think the math would work as described. The elo bonus won't be the seperate thing that sits there and accumilates, it will just be dumped into the current elo system and constantly redistrubuted. While increasing the elo pool would increase everyones elo some, it wouldn't increase it enough to fix the standings problems. I ran some numbers.

We gave it a try simulating it in simple way, but I can go and calculate it with exact numbers of this season (wins, losses, number of games) from beggining, simulating literally whole season just for you. As soon as I come on laptop

लिखा द्वारा Tungston., 05.02.2022 at 08:50

This is an interesting idea, but I don't think the math would work as described. The elo bonus won't be the seperate thing that sits there and accumilates, it will just be dumped into the current elo system and constantly redistrubuted. While increasing the elo pool would increase everyones elo some, it wouldn't increase it enough to fix the standings problems. I ran some numbers. Lets say we start juicing the current cw season with this elo, an extra 316 elo (games played this season) isn't going to push any of the current active clans mention above 1000 elo again, 316 elo isn't enough to do that i don't think. Can someone explain how the math would accomplish this, cause saying they played 50 cws and adding 50 elo isn't how it would work. ELO gets redistrubuted and wouldn't a lot of the extra elo just end up at the top with paradox and not really fix these standing problems? IDK, someone smarter than me figure this out.

I pulled every cw game played this season and calculated a "What If" leaderboard with a program, if we did the elo bonus system I said earlier.

What If


Here is the current leaderboard standing.


Important changes:
- Hellenism no longer last place behind inactive clan like constellation and backup
- Constellation and ReichsDNA booted from 4th and 5th to 8th and 9th
- Enigma is still behind "backup" but now the gap is ~4 elo instead of 39.8
- Prestige gap is a lot closer to beating backup. Gap used to be ~39 elo, now about 13 elo

Keep in mind that most games you win 12.5 elo so 2nd-4th place is only 1-3 wins away as opposed to literally 10+ wins for clans like hellenism now.

This is the data I've compiled and used for my program. It lists every single cw in the order they were played.
https://pastebin.com/FSjug95y

So if Mr.r̶a̶n̶ ̶m̶y̶ ̶o̶w̶n̶ ̶n̶u̶m̶b̶e̶r̶s̶ Tungston wants to go ahead and check if I made a mistake, he can feel free to do so.

लिखा द्वारा Witch-Doctor, 06.02.2022 at 11:14

लिखा द्वारा Tungston., 05.02.2022 at 08:50

This is an interesting idea, but I don't think the math would work as described. The elo bonus won't be the seperate thing that sits there and accumilates, it will just be dumped into the current elo system and constantly redistrubuted. While increasing the elo pool would increase everyones elo some, it wouldn't increase it enough to fix the standings problems. I ran some numbers. Lets say we start juicing the current cw season with this elo, an extra 316 elo (games played this season) isn't going to push any of the current active clans mention above 1000 elo again, 316 elo isn't enough to do that i don't think. Can someone explain how the math would accomplish this, cause saying they played 50 cws and adding 50 elo isn't how it would work. ELO gets redistrubuted and wouldn't a lot of the extra elo just end up at the top with paradox and not really fix these standing problems? IDK, someone smarter than me figure this out.

I pulled every cw game played this season and calculated a "What If" leaderboard with a program, if we did the elo bonus system I said earlier.

What If


Here is the current leaderboard standing.


Important changes:
- Hellenism no longer last place behind inactive clan like constellation and backup
- Constellation and ReichsDNA booted from 4th and 5th to 8th and 9th
- Enigma is still behind "backup" but now the gap is ~4 elo instead of 39.8
- Prestige gap is a lot closer to beating backup. Gap used to be ~39 elo, now about 13 elo

Keep in mind that most games you win 12.5 elo so 2nd-4th place is only 1-3 wins away as opposed to literally 10+ wins for clans like hellenism now.

This is the data I've compiled and used for my program. It lists every single cw in the order they were played.
https://pastebin.com/FSjug95y

So if Mr.r̶a̶n̶ ̶m̶y̶ ̶o̶w̶n̶ ̶n̶u̶m̶b̶e̶r̶s̶ Tungston wants to go ahead and check if I made a mistake, he can feel free to do so.

So much hostility at the end haha. I'm sorry WD for the personal offense you are feeling, although I''m not sure what that would be.If you would like to work it out feel free to PM me and I would love to make ammends for any personal offense you may be feeling.

Back to your post, cool idea and thanks for running the numbers. My calculations was definitely not as in depth as yours. Great work! My biggest concern I was unable to tell you yesterday was more a question as i did not know how the elo system worked as clearly as you do. My concern was the activity bonus would be deluted too much as the season went on and not boost everyone up enough to get past the 1 game cwers, but it looks like according to your calculations it would or would be close since backup is in 4th. But probably with another week of cws increasing the pool or a small win streak by a clan would probably fix that too. Again, Great Work!

I don't like cws, so much stress and bullying from clan mates.

(intro)
dj snoop dog on the beat

(verse)
they try to test us
motha fucka you mess with us
fuck with the G$ and you take up my trust
fuck you wall up after i raise up my cup
this is a battle cry bitch
cuff your hands bitch
you make some gay elo calculations?
the only thing you got is gay relations
pay attention bitch
this ya lesson bitch
there aint gone be no treaty
with a n*gga called WD
my hands jittery
shit cant wait to blast these
leave the n*gga wd bleedin
this shit just ez pz

(hook)
fuck out the way with this elo calculations
fuck out the way with this elo calculations
how bout you tell me bout your elo molestations
how bout you tell me bout your elo molestations

लिखा द्वारा Tungston., 05.02.2022 at 08:50

This is an interesting idea, but I don't think the math would work as described. The elo bonus won't be the seperate thing that sits there and accumilates, it will just be dumped into the current elo system and constantly redistrubuted. While increasing the elo pool would increase everyones elo some, it wouldn't increase it enough to fix the standings problems. I ran some numbers. Lets say we start juicing the current cw season with this elo, an extra 316 elo (games played this season) isn't going to push any of the current active clans mention above 1000 elo again, 316 elo isn't enough to do that i don't think. Can someone explain how the math would accomplish this, cause saying they played 50 cws and adding 50 elo isn't how it would work. ELO gets redistrubuted and wouldn't a lot of the extra elo just end up at the top with paradox and not really fix these standing problems? IDK, someone smarter than me figure this out.

This aged really well.

I actually put a lot of effort into my equation I think we should still use it

लिखा द्वारा Witch-Doctor, 03.06.2022 at 17:49

लिखा द्वारा Tungston., 05.02.2022 at 08:50

This is an interesting idea, but I don't think the math would work as described. The elo bonus won't be the seperate thing that sits there and accumilates, it will just be dumped into the current elo system and constantly redistrubuted. While increasing the elo pool would increase everyones elo some, it wouldn't increase it enough to fix the standings problems. I ran some numbers. Lets say we start juicing the current cw season with this elo, an extra 316 elo (games played this season) isn't going to push any of the current active clans mention above 1000 elo again, 316 elo isn't enough to do that i don't think. Can someone explain how the math would accomplish this, cause saying they played 50 cws and adding 50 elo isn't how it would work. ELO gets redistrubuted and wouldn't a lot of the extra elo just end up at the top with paradox and not really fix these standing problems? IDK, someone smarter than me figure this out.

This aged really well.

HAHA, Quotes pulled out of context rarely do age well, you are correct as usual!
Did you do this to get some compliments Well who am I to deny? Here you go bud

I've said it multiple times before and I'll say it again cause its truth and not just sarcasm. WD, this was a great idea. It worked great last season. Keep the good ideas coming. Cheers!

And in case that last message gets deleted as others have in the past, I'll say the compliment again.

I've said it multiple times before and I'll say it again cause its truth and not just sarcasm. WD, this was a great idea. It worked great last season. Keep the good ideas coming. Cheers!