As mentioned in other threads, Clan Wars is developing into a two-tier system.
Three clans have broken out as consistent top performers and are increasingly leaving behind the rest of the pack.

As such, The top 2 clans are the only ones competing over the Rank 1/2 awards, and everyone else seems to fall into place against the territory awards.

I've compiled a bit of data from the past 4 completed seasons along with the almost completed 5th season. Overall, the Duration of each Clan War is creeping upwards when scaled for the # of territories in the map.

Season                      1       2        3       4       5
Duration (days)            14      43       31      38      42*
Clans                      63      95       87      83      85
Territories               183     488      385     395     413
Terr/Day                 13.1    11.4     12.4    10.4     9.8
Avg Ter/Clan              2.9     5.1      4.4     4.8     4.9
 
Top 3 Clans
Terr/Day                  1.8     1.7      2.3     2.5     2.6
Avg Ter/Clan              8.3    24.3     23.3    31.3    35.7

Remaining Clans
Terr/Day                  11.3    9.7     10.2    7.9     7.3
Avg Ter/Clan              2.6     4.5      3.8    3.8     3.8

The progressive number of wins to reward a territory to a clan is in part contributing to this lengthening of the seasons.
With approximately 60 territories left in the Season 5 map, the season could extend another ~week, topping the previous record for a 43 day season.

The average territories/day has slipped from 13.1 in season 1 to a current (and expected to drop) 9.8 for season 5.

Except for the top 3 clans, the remainder of the clans are only getting 3-4 territories per season, no matter how long the season goes. Thus, half of the clans are not even getting half of the territory rewards available to them. Meanwhile, the average number of territories claimed by the top 3 guilds have skyrocketed from ~8 to ~36 over 5 seasons.

So is this a problem? Well, rewards are given on a per-season basis. So if you double the length of a season, you are essentially giving away half of the rewards. As rewards are reduced, less people will play, leading to longer and longer seasons, thus further driving the cycle of reduction in rewards.

As a proposed solution, the number of wins needed to capture a territory could be reduced such that the first 10-15 territories only need a set 3 wins per territory, then the progressive number of wins could implemented. This would provide further incentive for the lower level guilds to earn the first 6-7 territory rewards. This would thus raise the average number of territories awarded to the bulk of the guilds, allow more players to receive the territory reward incentives, and thus shorten the season cycle to get more people interested in playing clan wars.

In designing Clan Wars, a lot of effort was put into making sure that it’s fun for everyone who participates. In a traditional ladder, I feel it’s a lot of fun for the top 10%, who are competing for a high rank, but pretty bland for the majority.

     - https://www.warzone.com/blog/index.php/2021/03/update-5-10-clan-wars

In light of this statement about the design intent around CW rewards, it does seem that the growing gap b/w the top 3 & everyone else on getting good rewards should be a concern.

Does your data also suggest that participation in CW has been declining for non-top-3 clans? Also, how did you get this data? I wonder if we can use it for other analyses, e.g., to see the impact of CW on player migration to the top 3 clans. I've noticed the top 3 recruit high-performing players from non-top-3 clans pretty aggressively (intentionally or otherwise). Makes it much harder to climb up as a lower-performing clan because you probably depend on some players who participate heavily and play well but then the next season those players go to MASTERs, M'H, or TSFH.

Imo the nice rewards remain perpetually out of reach for non-top-3 clans under CW's incentive structure, and it's created this self-sustaining hierarchy where some clans have become basically magnets for high-participation players.

The top 3 even represent this niche quite well:
- High-participation and high-skill? Join MASTERs to get better rewards next season for your performance
- High-participation and medium-high-skill? Join M'Hunters to get better rewards next season for your performance
- High-participation and medium-skill? Join TSFH to get better rewards next season for your performance

Meanwhile, CL has a much more fluid ruling hierarchy and more barriers to the top clans cannibalizing top players from lower-performing clans who are incentivized to leave for vastly better rewards.

On that note, another suggestion: individual performance awards in CW. CW can be a team competition but that does not mean every aspect of CW (ratings, rewards, etc.) needs to be grouped at the team level. I've seen this in other games where the same contest has rewards for top team performances but also recognizes top individuals so they have an incentive to perform well independent of how well their teammates do.

Edited 9/2/2021 20:26:41

I went to each season, View Map, then counted the frequency of the numbers of territories. I also went to the "G" page of the winning clan and grabbed the start time of the 1st and last game for that clan and assumed it was the start day/end day of the season.

Regards to your inquiry about reduced playing, look at the avg terr/day for the top 3 clans and the rest of the pack...
For the top 3, its climbed fairly steadily from 1.8 to 2.6, whereas for the rest of the pack, its slowly dropping from season 2 to season 5.

So yes, I think participation is dropping for anyone outside of the top 3 guilds.

I see. I hope we're able to get more data like this because I think post-5.12 updates have fundamentally redefined WL clans (as a concept), now that in-game rewards have become a/the key motivator for joining clans.

has fizzer ever implemented any changes when the community has complained? or does he always make up a reason to prove he was correct

has fizzer ever implemented any changes when the community has complained? or does he always make up a reason to prove he was correct

He makes up a reason. For CW the reason is, that it is a new competition, so making any changes is unwise.

get more people interested in playing clan wars.

Anyone want to source (either manually or automatically [in a non-ToS-breaking way]) rich participation data (by time slot & season) for CW? If we know who played when, for which clan, and on what template, and had some other data about users, we could capture a much better picture of CW participation.

First season was very different in terms of territories, not sure if it can be compared to later seasons so easily.
Also, the number of wins needed for territories also different between the seasons. Like, for example, now we need 20 wins for territories, but it was 25 at the end of season 3 etc.

I would argue there are 3 tiers, rather than 2, but the middle one is not so constant, and is growing. Not sure if this is important to the point. It does hint the competition on some ranks is increasing, somewhat, maybe?

What would be better indicator imo, is the amount of games played, rather than territories. If I had the knowledge how to do so easily, I would actually compare the amount of games played daily, and how it changes throughout each season. My assumption being that people play more in the early days, then the participation slowly declines, and maybe it decreases sharply right at the end, when the chance to gain another territory or surpass another clan in standings is small (that is except tier 1 and maybe tier 2 clans too).

The question here really is "is participation decreasing? if yes where / when?". What generally should be a goal, is to somehow increase it on all ranks. Like l4v.r0v mentioned, right now active players actually move to tier 1 or tier 2 clans if possible, so participation in lower ranks may not increase at all. I don't think this is something that can be dealt with.

Not without better motivation for the players to play, regardless of rank. I did mention the rewards are really lackluster in some other thread; the main reason to participate for me personally are some (questionable) bragging rights, wanting to make clanmates happy, and making Seph not mad.


Nonetheless, Fizzer did reply somewhere that he is observing how it plays out for now, so I would not expect any significant changes soon.

PS: Last note on this: I think the average daily participation could be raised simply by having shorter seasons (less territories on the map). First season was really not too short, and it could make standings less predictable actually, so more exciting.

Great post Santa

If I had the knowledge how to do so easily, I would actually compare the amount of games played daily

Since CW games are automatically created, you can get successful responses for them from the query game API. Once you get that, you can keep tabs on when the games were played, on what template, by which players, for which clans, etc.

However, the hard part is:
- getting the list of games (no API endpoint): the simple solution is to traverse game IDs in a binary search style until you encounter a CW game (e.g., increment game ID by 10000, see if you overshot or undershot, then correct) and then traverse IDs around that CW game until you stop seeing CW games. This is brute-force but uses only Warzone APIs so it might be preferable for compliance reasons. A less brute-force solution would be to extract the game IDs from the time slot page (which will also simplify data aggregation), but this is possibly ToS-breaking (regarding the automation restrictions) and not straightforward since you need a logged-in client (at the very minimum a headless browser, but more realistically a userscript) and because UJS code is obfuscated so you can't just get the game IDs from the HTML on the timeslot page and afaict would actually have to simulate button clicks (unless there's something buried in the JS). Given the complexity of the non-brute-force pathway, I suspect the difficulty is intentional and a sign that this approach may be undesired data gathering (unclear; there's existing userscripts that gather data from UJS pages, so not sure what the cutoff is).

- accounting for players who signed up but did not get a game (you can't do this without going to the time slot page, and even then you have to cross-reference to figure out what clan they were playing for at the time)

Basically this is just enough of a hassle that I don't think anyone will do it. But it's a few hours of work at most. And once someone just gets the game ID list, that alone will unlock a wealth of data to analyze using the query game API.

Edited 9/3/2021 00:52:37

Being in one of the top 3 clans, take my opinion for what it's worth, but I find it fun albeit simultaneously a chore to fit into the timeslots while also finding a desirable template. While it would be fun to also challenge for ranks 1 or 2, I can't say we don't have the ability to do so, just lack the drive to make it happen. Meaning, WR aside, both #1 & #2 consistently have more members playing everyday. All of the top 3 have 40 people playing each season, but we don't have enough people playing actively to be able to make a challenge.

Now look from #4 and down, it's the same trend but worse. They typically have even less members playing consistently everyday. But it's also these same teams that claim they don't have a reasonable chance of competing for top spots for various. And without discounting those reasons, I would counter with, the first step is getting more members to play every day. If you can't at least achieve that, doesn't matter what else changes in CW, you'll never make a play for the top ranks.

That aside - I do think that the Territory rewards should be more attainable. The fact that only the top 3 clans make it to 20 territories for the elusive +15% AP bonus CW reward can certainly be a deterrent for activity. If the only territory rewards they're realistically going to get are the ones they want the least, it's deflating.

The idea of individual rewards in additional to group rewards is a great idea. As it stands now, either the whole clan succeeds or fails together, and this makes it difficult for an individual member to be motivated if the clan as a whole isn't already motivated. This is fine for the top 5-ish clans, but for every one below that, it means someone has to incite vigor in the ranks in order to even get more than 1-2 territories, and this is hard. At least if there was some _desirable_ (ie: not just some crappy reward, something players actually want and can use and is material) reward(s), then an individual can start to play, get benefit, and this can be motivating. It all starts with 1-2 members, and can spread to others. Maybe rewards for 10, then 20, then 30 games played in a season (played, not won), and a different reward for 10, 20, 30 wins per season. Add an MVP distinguishment for clan (most wins?) per season. Technically 1 FC is given to everyone player, so yes, an individual reward does already exist, but it's just for playing any game at all, and obviously it isn't enough of a factor, b/c there's still not enough activity.

As for Fizzer listening to the community - yes, he has listened a few times, but more often than not, no, he doesn't. The case has to be convincing, meaning you have to actually convince him it's for the better of the game. He won't just implement something b/c many people want it, that isn't enough. Times he's listened? When we complained about the cooldown of Active Artifacts - he lowered it from 24h to 16h (sure, it's still not enough, but it's an improvement). And when he reinstated the empty bonuses for superbonuses to support IDSS (whatever the Iskander whatever whatever whatever abbreviation is). And a few more. Yes, for each time he listened, there are 100 times he didn't, but he can't implement everything, and we shouldn't expect it (and let's be honest, some requested items are just plain dumb and should never be implemented anyhow; but certainly there are some gems and missed opportunities). And who knows, maybe all these times he did listen (except the IDSS, that was legit making a case to overturn something he materially disagreed with because he didn't understand why we wanted it, and once he did, he saw the value & changed his stance) it was only b/c it was already something on his list that he happened to come to the same conclusion as the community on. Anyhow, what he did say about CW, the TLDR is basically to wait another 6-10 months to see how things go and maybe they'll stabilize and get better. So don't expect any short term changes.

In designing Clan Wars, a lot of effort was put into making sure that it’s fun for everyone who participates. In a traditional ladder, I feel it’s a lot of fun for the top 10%, who are competing for a high rank, but pretty bland for the majority.

And that's the problem, isn't it? The top 5% are happy, the next 5% are happy-ish, the next 80% are unhappy/complaining that they don't have more opportunity. So it's not really accomplishing the 100%-happy-across-the-board goal that it's aiming for. But it could ... and the community has provided lots of options to improve it.

And lastly ... I've said it before and will say it again - the timeslots are terrible, the template selections are terrible. Basing one's routine schedule around 4h timeslots is soooo inconvenient, and exacerbated by being online at the right time and finding no preferred templates. If the goal is widespread participation, this seriously needs to change. No idea why it's so fixated on fixed 4h timeslots and including templates that numerous people have indicated they don't like. This deters participation. Make it easy to find (A) a time to play, and (B) a desirable template. There are many options for this so I won't even go into a discussion.

These posts are well thought out. I just wanted to point out that two of the top 3 have 80+ members. So if all clans only have 50% participation then they have 40 members doing clan wars. Where my newer clan is limited to 40 members and would only have 20 members doing clan wars. Newer clans cannot realistically compete with clans twice their size.

@navrunner just remember that the large size of it doesn't matter as much since they cap the amount of participants for CW per clan at 40.

There is obviously a hierarchy in ranking between clans (Clear 1-3) that are extremely active and good for Clan Wars in specific, and thus I don't think anyone is really complaining for the Rank Rewards. The main argument being made here is that the later Territory Rewards are simply too impossible (unrealistic) to reach for every other clan except the Top 3. (The biggest one I have my eye on is the +15% AP which is the 20 territory reward. Season 5 seems to be the first time outside of the Top 3 to finally have 2 more clans in sight (Excel/Polish Eagles) to achieve the 15 territory mark giving the 20% dig cd reduction.)

Anyway though, if nothing is done in regards to this, many can (and already do) just simply view these territory rewards as rank rewards themselves for the top 3 basically. "You might as well just make those Territory Rewards exclusive to the Rank Rewards, because the odds of any other clan getting to them are extremely slim" (The only way this can change is if several of the less active clans fuse their active members, but I don't think it's likely for this to happen, and if it did, maybe 1 or 2 clans would get the ability/incentive to do so)

@navrunner just remember that the large size of it doesn't matter as much since they cap the amount of participants for CW per clan at 40.

Pre-5.12 clans (without the 40-cap) can hit the CW participation cap at 50% participation or less. For a post-5.12 40-capped clan to hit full participation, they need 100% participation. There's no way for post-5.12 clans to follow krinid's simple suggestion without extraordinary effort and churn (because they can't afford to have a single non-participating member in the clan while TSFH can keep around non-participants).

To date, the justification for the 40-cap at the clan level (because CW has a 40-cap) has remained a non-sequitur, just like the rationale for matchmaking CW off a single incredibly noisy clan-level rating ("because it's a clan competition"). It's fundamentally hard when every design decision in first-party community events remains constrained by needing to convince one evidently stubborn authority.

---

If the Create Game API's restriction were removed + a platform was made available where community events could broadly reach players including invisible casuals and newbies (like how QM and CW are highly visible to everyone), I bet we'd get something drastically better, more enjoyable, and far more responsive to community feedback than Clan Wars in its present state.

Perhaps we should look at the big picture: the successful, enjoyable, and continuously-improving (rather than degrading) community events have all been community-made, community-led, and community-owned. But the community has been constrained and shut out- the APIs have restrictions to prevent theoretical abuse scenarios (which would be trivial to detect if they occurred at meaningful scale) which make them borderline useless and further insulate the strategic community from the broader playerbase.

This is not the first thread to discuss CW design, and no one seems optimistic about any of this thread's suggestions getting taken seriously. Farah has had an entire class of students working on data gathering and analysis to make the case for a clear-cut obvious change on the 1 v 1 Ladder. CW and QM have fixed templates that, based on relative CW join rates, look to be empirically less popular than the templates developed and tested organically by the playerbase even though the fixed templates (by virtue of cluttering the initial unlocks on QM) have an edge when it comes to visibility and adoption. Meanwhile, MTL solved its rating system problem years ago, player-run events have competitive mechanisms that favor templates players really enjoy, and Clan League improves each season (in spite of the considerable effort it requires just to keep the lights on).

The broader case we should be making is for an ecosystem that promotes rather than handicaps community events, because the group that's most in-touch with the playerbase is the playerbase itself. We need more than half-functional APIs and a half-maintained Community Events page.

Edited 9/3/2021 04:10:12

@navrunner
Some great points. I agree that pre-40 cap clans have an advantage, but I think it's a smaller advantage than people think. Most of the 40-cap clans that aren't doing well simply don't have players joining games often enough. So using the logic of 50% CW participation rate, that puts 20 players in scope, and if a clan had 20 players playing at least every other day, they would likely make the top 10, maybe even higher. But they don't, many of these clans don't even have 20 players participating, nevermind active players.

But don't think that new clans can't be competitive - they can, look at Excel in 5th place with 40 members, Rep of Korea in 8th place with only 29 members, Python in 9th place (with only 36 members!), but also consider that Masters at #1 spot only has 54 members. So yes, >40 but not 80, 90, 100, etc, just a mere 14 beyond what the new clans have. They've optimized to be CW competitive.

Cyrkon has 40 members, but only 17 CW players, not even 50% participation. Even among those 17, only 3 have played >10 games. What this says is that the clan as a whole isn't serious about CW - which is fine, but if members want to be competitive, they either have to incite consistent activity, or leave the clan for one that has activity.

The unfortunate part about 40-cap is that it forces new clans to re-organize if they want to be competitive. Old clans still need to recruit new active CW players, but new clans need to both recruit new and cut non-active CW players. It kind of forces you to decide between being a general community clan or a CW-focused clan. This is the advantage that the old clans have.

So using the logic of 50% CW participation rate, that puts 20 players in scope, and if a clan had 20 players playing at least every other day, they would likely make the top 10, maybe even higher. But they don't, many of these clans don't even have 20 players participating, nevermind active players.

It's a vicious cycle and uphill battle because 40-capped clans have a far weaker incentive for players to participate (no chance of getting the nice rewards) and struggle to retain talent when breakout players hop to pre-5.12 clans that offer better opportunities for CW powerhouses.

Excel in 5th place with 40 members, Rep of Korea in 8th place with only 29 members, Python in 9th place (with only 36 members!), but also consider that Masters at #1 spot only has 54 members

Excel, RoK, etc., are likely one-season wonders. Look at the past 40-capped clans that have had strong CW seasons. They fade the next season because it takes a lot more effort to get them near the top than it does to get a fraction of TSFH to participate. Python is an edge case as one of the few clans that can get >50% win rates reliably in CW.

The unfortunate part about 40-cap is that it forces new clans to re-organize if they want to be competitive. Old clans still need to recruit new active CW players, but new clans need to both recruit new and cut non-active CW players. It kind of forces you to decide between being a general community clan or a CW-focused clan. This is the advantage that the old clans have.

And it's a tremendous one. If Ursus has school or work that makes him less available for one CW season, you're not forced to kick him out (and can benefit from his participation in the next season).

Unfortunately, due to the built-in difficulty of fetching data about CW, it's a pain to get numbers here (so it's a hassle to look into things when the admin jumps here and cites some non-auditable cursory data to call players' claim "nonsense"). But if you looked at which clans have lasting CW success, I bet it'd paint a different picture than the 40-cap being only a minor handicap.

Remember PKU58ers? Or look at the current best 40-capped performer in CW: KING. KING lost its two best historical performers to TSFH and their steady improvement more or less comes from Fausto somehow managing to marshal steady participation while the other non-top-3 clans in CW face dwindling results.

Edited 9/3/2021 04:20:08

Big 40+ and definitely have a huge advantage, as knyte said. There's no doubt about that. And that's coming from someone in one of the bigger clans.

40+ clans do have an advantage, one problem which might affect both the small and big clans is a random troll/member joining for one game and never playing again.

I don't expect most of the players who participate to know about this. so when someone participates for fun once that slot is gone for the whole season.
now this is even more problematic for the newer clans who use the "open clan" feature. I have seen trolls join, play and then leave. maybe we can think about a solution for this?

Edited 9/3/2021 16:25:35

As to the troll problem, perhaps only the top 40 slots are used to earn territories? But is that that big of a problem to make a change? Maybe a clan leader could permit members to play CW or not.

Maybe a clan leader could permit members to play CW or not.

Well, then you get into the problem of big clans effectively creating a roster/lineup thing for CW. The whole point of CW is that its for everyone in the community, not just the good players from decent and better clans.

Also, on the topic of getting gameID's for CW games, I hope Fizzer adds that as an option to the GameFeed API. That would solve a lot of problems and allow for a much better analyses of CW.