Data for the current matching system for Clan Wars.

Edited for content on 1 July
Hypothesis: the matching system, assigns matches, and the weighting factors really don't change across the season. In cases where all top five clans are queue for a template The number one clan is or always first matched with the number four clan. There is no randomness in the system.

History: In response to CWR *manipulation* and numerous complaints about the matching system, and *gaming* of the Clan War Rating to improve matchups, Fizzer took action. He said that teams should play those ranked near to them, that the matching algorithm would be modified to make this happen. This was implemented with the September 2025 version, adding a "match rating" factor and noted in this blog. https://war.app/blog/index.php/2025/09/

• All: When matchmaking, clan wars now blends your clan’s war rating with your clan’s current position in the leaderboard for that season.

So the question is does the system need more tweaking?
Is the matchup manipulation problem fixed now?

Lets do some analysis in the next post.

Edited 2026-07-01 18:13:53

Lets see what the matching system does. Taking a few examples.

https://war.app/Clans/War?ID=82&Timeslot=11759

Template SE1W

1st Clan. 3 players
2nd Clan. 3 players
3rd Clan. 2 players
4th Clan. 3 players
5th Clan. 1 player

Matchups for 1st Clan

First clan skips over **five** third and fourth place clan players to play clans ranked 4th and 13th.

Template: Med Earth Trad.
Also for Medium Earth Trad, the 2nd place team member was given the free win, while 1st ranked clans played clans ranked 4th and 8th.


https://war.app/Clans/War?ID=82&Timeslot=11743

SEAD template

1st Clan. 2 players
2nd Clan. 6 players
3rd Clan. 10 players
4th Clan. 4 players
5th Clan. 4 players

Matchups for 1st Clan

1st Clan skips over **sixteen** 2nd and 3rd place teams to play clans ranked 4th and 5th.

Also in that same timeslot, in the bomberman template, 1st place is matched up with 8th and 5th place clans, skipping over higher ranked clans.

We could do this for hours.

The current clan wars matching system is frequently skipping over ranks in matching teams. Or does it? I could not find many counter examples where alll top 5 clans were in a template. Certainly, if there are counter examples, then someone will find them. But, there is one more factor to discuss in the next post.

Edited 2026-06-30 14:41:00

Edited for content July 1

The evidence is the CWR manipulation by some clans at end of season. Look at the last-day games for this clan, for example: over the last two seasons:

Season 81: https://war.app/Clans/War?ID=81&ClanGames=345
Day 21 Games: 20 and points: 5
Day 20 Games: 20 and points: 3.

Season 80: https://war.app/Clans/War?ID=80&ClanGames=345
Day 27 Games: 20, and Points 1

A quick review of the "lost" games shows quite a few insta-surrenders.

So, by at least 1 Clan, CWR manipulation is a real thing, and is giving them "easier" matchups. Hats off to them and they are working the system in its current form.

---

CONCLUSION:

Current matching algorithm always pairs -- in the current cwr environment-- clan one with clan four, when top five clans are in the template. The effect of the clan rank is not apparent., matchups do not change much over the course of a season. With the current CWR and rank values, clan one is always match with clan four, first.

Recommendation: Tweaking the formula to more randomly to favor the Clan Rank over the CWR, will lead to, at least some, matchups between the clans ranked in their order, as opposed to skipping over clans.

Edited 2026-07-01 19:46:25

Timeslot 2 of 6 today

https://war.app/Clans/War?ID=82&Timeslot=11762

SEAD template

1st Clan. 3 players
2nd Clan. 4 players
3rd Clan. 3 players
4th Clan. 3 players
5th Clan. 1 players

In this template, All Three of the top clan were matched with the 4th clan. They skipped over **seven** players in clan 2 and 3 ranks, to play the 4th ranked clan. The 2nd and 3rd place clans all had "easier" games.

The pattern is noticeable.

Edited 2026-07-01 19:46:31

I think you're making too big a deal of it, but I agree that CWR manipulation is still a thing. The regular quick surrenders at the end of a season are evidence of that.

That said, I don't think it makes much of a difference anymore. That was the point in Fizzer's changes. Sure, knyte lowers their CWR by maybe 100 at the end of a season, but it is quickly raised in the first few days and doesn't change the outcome at all .

I personally would prefer intentional losing to be heavily dissuaded, as I would in any sports. If someone shows up to have a fun game, they should have one.

This isn't like the NFL or MLB where there's a desire to have competitive games. The system is set up where it's pretty natural for the best players to migrate to the top team (MB) to get the best rewards. Right now the second best team tries all the tactics and system gaming it can to make a difference but it doesn't make a difference. Is that a serious problem?

knyte club is lame and Prime/Harmony are both pretty cool

I think any CWR manipulation benefit which knyte club get by throwing games is offset by the reputation impact which comes from most players thinking they are fucking losers

Incidentally, these frequent 1000+ word, multi post tirades don’t help Harmony in this regard

Continuing to document the matchups.

https://war.app/Clans/War?ID=82&Timeslot=11765
SEAD Template
Here the lone 1st Clan Player skipped over Eleven 2nd and 3rd place clans to play the lone 4th place clan.

It's Every Time.

Edited 2026-06-29 16:21:24

Easy Fix
======

So in conclusion, the matching algorithm might benefit from additional tweaking.

What I suggest is to add a random variable to the matching system.

A Bernoulli random variable, with a 50% chance of choosing "closest CWR," and a 50% chance of choosing "closest rank," meaning half the matchups to be ranked based and half to be CWR based. This could easily be back-tested by running the revised matching algorithm against any of the timeslot cited in this forum thread.

Recognize it doesn't stop anyone from manipulating their CWR and getting rewarded for it. It's just not 100% of the time.
When the whole clan comes together to do some thing I have to applaud it -- your CWR manipulation is sublime. Hearty Handshake.



I don't speak for Harmony. I'm just publishing some data here for public consumption. Any views are just my own, they don't represent Harmony. I've gone back and edited the original post, above, to align more with the data.

Data analytics is my religion. I'll leave the politicking to other people.

Edited 2026-07-03 20:23:02