We are at a point when clan wars are so unfair that just matching people completely randomly would be better. The current system has multiple issues:
1) not being able to find matches even when there are enough players
2) separating clans into sub groups where some clans never face each other and some clans always fight each other
3) creating players with scores like 0-20 in clan wars
4) unfairly giving free wins to the same clans over and over again (Harmony and Ving being the worst examples here)
5) enables meta strategies like time slot map swarming and free win hunting

The system needs to be changed either to use per player CWR instead of per clan one or a fully random matchmaking to make things more fair. If the current system had an aim to make clan wars fair, it fails at that both at individual level and on the clan level. If the aim was to find matches for everyone, it fails at that as well, as we clearly see timeslots with enough players to find matches for everyone and 80% of a single clans roster being left out regardless.

This was the TLDR version, I will go into detail for each point:
1) not being able to find matches even when there are enough players
This is happening when mid CWR clans enter a map in a timeslot with a large number of players. Let's assume we have the following clans:
A - 800 CWR - 5 players
B - 750 CWR - 4 players
C - 600 CWR - 9 players
D - 300 CWR - 1 player
In this scenario the matchmaker will make 4 games between A and B, 1 game between A and C and 1 game between C and D. Then, 1 player from C will get a free win and 6 players from C won't get a match at all. A more fair system would be to pick 1 random player from the 19 available players, give them a free win and then match the remaining 18 players into 9 games. If C doesn't get the free win, that would mean every single game would be between C and some other clans, and if they do get a free win then 1 game could be 2 other clans facing eachother and clan C would be playing the rest. If we stick to the current clan CWR, then it would be 1 game between A and B, and the rest would be C against everyone.

2) separating clans into sub groups where some clans never face each other and some clans always fight each other
This is completely bad because we are not playing the same clan wars at all. How can we compete for the same reward while having matches of different difficulty? I know for a fact that if I go into an S1V1, I will fight OP 80% of the time and if I go into SEAD, I will fight Prime 80% of the time. A clan like ours and Prime now have to fight each other timeslot after timeslot while Harmony will fight whoever is with the lowest CWR in the timeslot. This then compounds because if you win against a higher CWR clan, you get more CWR and if you win against a low CWR clan, you get less, therefore making the whole system inert and more unfair than random selection. If we are competing for the same territories, we should be competing against each-other, not small islands competing within themselves.

3) creating players with scores like 0-20 in clan wars
This is probably the worst part of the current matchmaking. If you have clans with wide skill distribution, some players will end up with insanely high and some with insanely low win rates. Let's say you have a player with 800 skill (whatever that means), and one with 200 skill. Since the 800 skill player will keep winning and the 200 skill player will keep losing, their clan CWR will not move. However, the 200 skill player will lose almost all of their games. This is very unfun, if you want fairness, you need to aim to keep people between 70%-30% win rates. A single number for an entire clan makes this completely impossible. Either go full random, or per player CWR. I would personally prefer full random to avoid creating player islands similar to the clan islands we have now, or a weighted random, where players with similar personal CWR would be more likely matched, but not matched to each-other by force.

4) unfairly giving free wins to the same clans over and over again (Harmony and Ving being the worst examples here)
This has the same explanation as #1, except in this case it's about the clans on the lower end of the CWR. The free win system is predictable and easy to farm if you have low CWR and it does nothing if your CWR is mid-high. This in theory shouldn't be a problem because you are giving free wins to the weakest clans. However in practice #3 clan has #1 most free wins. A player CWR based system or a random system wouldn't have this problem by default.

5) enables meta strategies like timeslot map swarming and free win hunting
I have already expanded on free win hunting, so I will explain timeslot map swarming here. Having all the players from a clan in the same timeslot currently games the system because you know that there will be N games available against a stronger clan against you. You don't know N, and you have no effect of it. However if you want to get easier games you need to have your number of players higher than that N. Let me show with an example what I am talking about:
clan A - 800 CWR - 5 players
clan B - 750 CWR - 3 players
clan C - 700 CWR --- YOUR CLAN
clan D - 650 CWR - 8 players
In this scenario, 3 players from A and B will be matched, and the N I was talking about will be 2. Therefore if your clan goes in with 1 or 2 players into the map, they will be all paired with players from clan A. But if they come in with 10 players, only 2 will be paired against clan A and 8 against clan D, Therefore the more players you have on the same timeslot, the weaker your average opponent becomes. As an extra, if you go in with 11 players, you will get 8 easier matches and 1 free win as a bonus.

What am I proposing?
1) Either a per player CWR score based matchmaking with weighted randomness or fully random matchmaking
2) Limit the number of players for the clan with highest number of players on a map in a timeslot in the following way:
A - number of players from all other clans added up
B - number of players from the clan with second highest number of players
Limit = Max(Round(2 * A / 3) + 1, B + 1);

If Fizzer doesn't have the time or doesn't want to put in the effort, I volunteer to write the whole matchmaker in whatever language he wants me to.

Edited 9/23/2023 22:51:54

4) unfairly giving free wins to the same clans over and over again (Harmony and Ving being the worst examples here)

I guess, that explains why Eternity and Harmony team were so after NONO (and his FTW/GN clans) ...hardly a coincidence
#BRINGBACKNONO
#GIVEFREEWINSTOPRIME

Edited 9/23/2023 22:45:39

Step by step to your 5 bullet points:
1 -> agree. straight forward agree. honestly solid idea cant imagine rn ups and downs of it. I have been more of the opinion that people should have 2 votes in a cw slot tbh. And in case of an even amount of players getting a game in your 2nd template. But this will never be implemented

2 -> that is true only for sead and maybe se1w, where the big number of contestants puts the people in an order of cwr and repetitiveness is the issue. but i see your point that a clan in the range of 600ish cwr will usually only face players of the same strength. that is however one of the strengths - pros of this system, as mb, op or any other clan with clearly bigger cwr will not just get random opponents and win everytime. CWR originates from sum of average player strength in a clan, and in a scenario where we all faced random opponents in random templates, the clan with the best players, in average, would win CW, by definition the clan with highest cwr.

I agree with you that you will preemptively usually know what to expect in each template, s1v1 op/mb, sead prime, etc. But then again this falls into your 5th point of clans employing stacking and other strategies for better matchups. And people ofc playing the template they are most comfortable with (there are exceptions here which you haven't considered).

I dont think the matchmaking system in this regard is perfect but it is consistent. if the matches were random, the better player would always win in the end. now it leans on rewarding participation and presence accompanied by skills and total wins

3-> i dont understand this at all. w/l ratio is after all representative of skill. whats the correlation between w/l and the current system and how a different system would benefit people of less skill to perform seemingly, on paper, better. I dont think it is a bad feature that players from the same clan have a diverse background in capabilities, with some consistently performing better than others. Rather, it promotes rivalry to improve. In harmony at least it does, our cw leaderboards are cool :)

4-> unfair is a very strong word to use. Cause you just open fivesmith's site, see harmony have 19 or whatever it is now free wins and blame the system. but hey take a look at that
ng = no game
https://www.warzone.com/Clans/War?ID=37&Timeslot=5562 1 ng 6th slot
https://www.warzone.com/Clans/War?ID=37&Timeslot=5543 1 ng 5th slot
https://www.warzone.com/Clans/War?ID=37&Timeslot=5538 2 ng 6th slot
https://www.warzone.com/Clans/War?ID=37&Timeslot=5526 5 ng 6th slot
https://www.warzone.com/Clans/War?ID=37&Timeslot=5525 2 ng 5th slot

and thats only harmony players. i cannot talk about Ving, i'm in Harmony. You see Harmony benefiting from free wins by seeing a number when the iceberg is much bigger than that. If you want to stop the free wins you can try and read the strategy behind the scenes. Secret tip most of them originate from playing in harder templates where people fear playing. But regardless, you see 19 free wins, I see with after a search of 3 minutes in https://wz-cw.5smith.ru/timeslots_list 11 no games, and every day during mostly 5th and 6th slot in Harmony chat we urge people to spread and play in different templates so as to avoid no game, despite not being a player's expertise. While fw are indeed an upside for a clan with relatively low cwr, there are downsides. All in all no games and free wins are equally unfair in my opinion as well, just let me be clear, again,

there are 2 type of fw. Ones that originate from lower cwr, which are risky and sometimes end up in no games, ending up in harming the clan more often that not. And others that originate from harder templates that other player deliberately avoid. Which is a skill and reading the moment issue.

5 -> 100% agree. every clan is employing different strategies to maximise their wins. Well maybe not every every clan out there, but at least the top4/5


For your propositions i fear the completely random matchmaking would go into favouring only skill and not participation which is very different from the current system. Would i like that? Yes personally, but it seems to me that Fizzer/Warzone are not striving for that with CW. But there are def potential improvements and a per player CWR score could work. Don't have other ideas, tbh haven't really thought more of this cause i dont see it ever changing

1 game between A and C and 1 game between C and D. Then, 1 player from C will get a free win and 6 players from C won't get a match at all.

thats not how free wins work, if any player from C gets a game, no-one else from C can get a free win

pros of this system, as mb, op or any other clan with clearly bigger cwr will not just get random opponents and win everytime.

That's not a pro, that's a con. How is your clan competitive with OP when you can't fight them? You are competing for the same rewards, if you want to be #2, you should be able to defeat the current #2. Also, even in a fully random system, they would get each other quite often. But that's why I said a weighted random system that favors players with similar CWR should be also ok.

CWR originates from sum of average player strength in a clan, and in a scenario where we all faced random opponents in random templates, the clan with the best players, in average, would win CW, by definition the clan with highest cwr.

And that would be bad, why? Players with higher skill should win more. Also, CWR doesn't represent anything. If I win against OP, my clan gets more CWR then when someone from Harmony wins against unknown_clan_495. Then multiple people from my clan need to lose just to get to the CWR we had before my unfortunate win.

if the matches were random, the better player would always win in the end.

No, they would randomly get matched against players that are even better. Plus I did say we can use a weighted random system that favors players with similar skill level. Also, the current CWR is just a number that does not represent the skill level of the given player in any way.

I dont think the matchmaking system in this regard is perfect but it is consistent.

That consistency is the major problem. It forces some clans to constantly face stronger opponents while other clans constantly face weaker ones.

i dont understand this at all. w/l ratio is after all representative of skill. whats the correlation between w/l and the current system

CWR does not represnt w/l ratio at all. It also takes into consideration against who did you win. Which is why your clan with higher win % have lower CWR than my clan for example. That's nonsense. You get easier opponents and then if I win against a harder one, I get punished for it by a huge amount of CWR increase for my clan that then forces my clan to face even more strong opponents. If we are fighting for the same rewards, we should have the same difficulty. Why would you fight a 5 year old kid while I am fighting Tyson in his prime and then get the exact same reward if we win?

how a different system would benefit people of less skill to perform seemingly, on paper, better.

Because in the same clan you have people with different skill levels. If you give them all the same points, the weaker players in the clan will always lose and the stronger ones will always win. It's a zero sum game where the weaker players get more and more demotivated.

Cause you just open fivesmith's site, see harmony have 19 or whatever it is now free wins and blame the system. but hey take a look at that
ng = no game

Point #1 was exactly about the NG problem. Both issues are caused by the same problem. Also, your 21 free wins against our 0 or Prime's 3 does give you a huge advantage. It's the reason you are in fact ahead of Prime right now.

For your propositions i fear the completely random matchmaking would go into favouring only skill and not participation which is very different from the current system.

A weighted random system that would match you more often against someone of your own skill and a system that has a per player CWR instead of per clan CWR would still somewhat reward participation, while also rewarding skill. Plus it wouldn't result in clan mates with 0-20 stopping CW altogether because they got demotivated by constantly losing.

thats not how free wins work, if any player from C gets a game, no-one else from C can get a free win

You seem to be correct, that explains some timeslots that I couldn't explain myself. Still it doesn't change the fact that the matchmaker still has all the issues from all the 5 points in the original post.

Edited 9/24/2023 00:41:32

so the witches did the witch hunt...
interesting now...

Another thing that I remembered, Fizzer actually has access to all the past CW games and if he runs an elo calculation on all of those games, he can get a personal CWR for every single player quite quickly. Then he can switch over from clan CWR to personal CWR and from fully enforced CWR matching to a weighted random CWR matching where you have the highest chance to match a person closest to your CWR and lowest chance to match someone who has very different CWR than your own.

i was going to write about exactly the same thing.
clan ratings for matchups needs to be abolished
i’m for a completely random system, that way it’s harder to hunt free wins also

Edited 9/24/2023 01:19:00

While I agree with some issues, which were highlighted (no-game issue, and to lesser extent the same-opps issues for top clans), the other ones ("unfair free wins", "0-20") just show the inability to adapt to the system and blaming it for everything.

And the suggestion of random matchmaking is just plain unsuccessful. It almost destroys two of the principles, that Fizzer stated for CW (and I personally agree with him in those):
- CW is a competition of teams, not individuals
- High quality matchmaking for top clans is prioritized, even if brings inconveniences for bottom-clans

There may be some "middle of the road" solutions to tackle some of the mentioned issues, but random matchmaking is definitely not good. I can go into the details how random or individual matchmaking brings a whole set of new problems, but would like to focus on principles for now.

On a constructive side, if you want your matchmaking been seriously reviewed by FIzzer, do some testing on either artificial or real CW data and present results, how that would change matches for everyone. That's what Fizzer recommended to do during one of the recent AMAs, when he was asked about changing the matchmaking.

the system is too easy to abuse. i’ll give a full breakdown of my experience when i have access to a keyboard.

The fact that NONO could even rig CW in the first place means there's something wrong

The facts that "the system is too easy to abuse" and "NONO could even rig CW in the first place" means that other clans can do that too (ofc, remaining within the rules). And if other clans don't do that to improve their results, means that either it is not that easy, or clans just don't have the will to compete.

The facts that "the system is too easy to abuse" and "NONO could even rig CW in the first place" means that other clans can do that too (ofc, remaining within the rules). And if other clans don't do that to improve their results, means that either it is not that easy, or clans just don't have the will to compete.

We all know how to do it.
Step 1: Lower your CWR by losing a lot
Step 2: Get your weaker players to go to unpopular maps 1 by 1 (1 player to 1 map) and the stronger players go together to the most popular map of the timeslot.
Step 3: If your CWR gets too high, lose against the weakest opponents on purpose, also do not win against any clan who has above 500 CWR ever.
It just feels cheap and borderline cheating.

On a constructive side, if you want your matchmaking been seriously reviewed by FIzzer, do some testing on either artificial or real CW data and present results, how that would change matches for everyone. That's what Fizzer recommended to do during one of the recent AMAs, when he was asked about changing the matchmaking.

Since you already did all the data digging, can we talk in private about this? I can write a code that gets every single timeslot and every single game, but I don't think I want to abuse the server that much if you already have enough data.

Edited 9/24/2023 01:40:21

Can confirm. i am one of the weak players and texx sends me to play guiroma and lose on purpose

We all know how to do it.
...
It just feels cheap and borderline cheating.

Yes, exactly. When looked in detail, it appears to be totally legitimate, but does not satisfy some moral standards, which every person has their own. And I would argue, that personal tastes do not justify "at a point when clan wars are so unfair that just matching people completely randomly would be better".

Objectively speaking, when multiple clans will try to do that, the effectiveness of that whole scheme for a single participating clan will get drastically reduced.

Objectively speaking, when multiple clans will try to do that, the effectiveness of that whole scheme for a single participating clan will get drastically reduced.

It would be funny to see multiple clans racing to the Surrender button to try to tank their CWR,

LMAO @ potential 50/50 shot at accidentally winning by enemy surrendering quicker than you can

And I would argue, that personal tastes do not justify "at a point when clan wars are so unfair that just matching people completely randomly would be better".

I am actually more for a weighted random system. In fact I do have a system that is quite simple but has both randomness and fairness to keep players within 30%-70% win rates and avoids all of the issues mentioned in the original post.

In fact I do have a system that is quite simple but has both randomness and fairness to keep players within 30%-70% win rates and avoids all of the issues mentioned in the original post.

As soon as there is a system that, no matter how bad I play, will keep my winrate within 30%-70% win rates, I will be super happy. If it will make my winrate 50%, that will be all I want.
(Now I need also someone to make a userscript shortcut, to press "run+commit" button of the autopilot).
I hope, no one will see that system as "unfair".

As soon as there is a system that, no matter how bad I play, will keep my winrate within 30%-70% win rates, I will be super happy. If it will make my winrate 50%, that will be all I want.

That is pretty much any matchmaking system in any competitive game. Even the current matchmaker tries to drag clan winrates as close to 50% as it can. Unless you are matching people randomly, the matchmaker will drag players towards 50% win rate because the closer it approximates your skill level the closer your win ratio gets to 50%.

Woah, those are long texts. I knew CW was going to be more important than CL (what even is CL? You guys don't care.). But boy, I did not expect that.

This thread already almost has as many replies as the CL 17 sign up thread. And people were actually reading the opening post and dissecting it.