How does CW sort within clans now? Is it random? I remember it's no longer based on your account ID.

What if that were just based on # of wins so far that season?

Yes, within a clan it was changed to random

1. Free win hunting is a design principle of the current system and it’s arguably a good one

Free win hunting is desired behavior. Warlight is a small game with a small player base. Fizzer has been very clever to strike the balance between timeslots per day + templates per slot to maximize the number of populated slots. An alternate universe where FWs are eliminated completely (i.e. you get a message that says sorry try again next slot) yields fewer templates options per day.

2. Pairing top clans against top clans is best practice

Simply put, the seasonal title should be decided by the top clans going head to head. Yes, using an individual player Elo would be desired but given the alt situation in CW today this is simply not realistic.

The problem with the current matchmaking system (in my mind) is specifically the situation where it appears there are enough players to find a game but in reality there are not.

To solve this problem, clan names (but not player names) could be revealed during the xx:00-xx:10 period, either as a list or as projected pairings so players have visibility when this may be a problem.

Alternatively, you could pair in such a way that the low-CWR / high-volume clans with a long tail which doesn’t get paired in the current format is instead paired first. This has the undesired effect of pairing top clans against bottom clans even when top-top pairings are possible so it’s a double-edged option.

SEAD is newbie friendly enough

A bit offtopic, but anything with 1 minute boottime is absolutely terrible for beginners. I stopped playing SEAD and SE1W solely because it's simply unplayable for me. Had this been my first multiplayer experience (instead of 5 min boot autogames on ME), I'd have left this site immediately, instead of sticking around for over a decade.

Yeah if there's anything newbies hate it's when games move fast and end quickly

I can't believe people are trying to fix a system where the best performance is measured by NUMBER OF WINS.
This is not possible by design.
Later in the thread, it gets even funnier, when OP slowly seems to realize that rating is a better way to determine who's best. And fire is hot.

that's the problem tho, for new players, it can be too quick, for veterans (not necessarily elites), it is often too slow

in the context of CW, many people play SEAD not b/c they actually like it, but simply b/c it's the template with the lowest timers and is often the quickest games - not even a picking phase, so it saves time there too

ultimately what clan wars needs is better competition between clans. everyone, except mb was tired of them winning every season. opti made some moves and now there’s a great race. sane between harmony and prime. harmony is able to motivate members when the placements are close and there’s something to play for. i’m sure prime is putting a great effort not to lose 3rd place.
these battles are what drive the players in the end.
more than individual matchups, better equality between clans is the motivation factor for player retention imo. harmony has been focusing on and pushing participation above all else since last year, it’s taken a while but it’s finally bearing fruit.
i’m not a fan of the current matching system but that’s not the big problem.
i’m old yeller and thanks for coming to my ted talk.

for the record, i applaud what mb and opti do. prime has been a top cw clan for seasons also. it’s up to the rest to figure out how to catch up.

Edited 9/27/2023 18:51:27

0%-20% winrate - 69.93% stopped playing
20%-40% winrate - 59.81% stopped playing
40%-60% winrate - 58.11% stopped playing
60%-80% winrate - 60.29% stopped playing
80%-100% winrate - 69.04% stopped playing

Interesting. Let me make a suggestion based on that.

People who are active Idle-players are prioritized in matchmaking and we give them 40-60% matchups based on individual rating. The other players can have the remaining matchups based on whatever algorithm.
This will increase satisfaction of idle-players, and CW is CLEARLY made for idle-players. We want them to be happy customers.

Being one of those idle players, I have to say that I am absolutely more than fine with the current match making and would argue that the no game situation is really so irrelevantly low, that it is not worth introducing other abusing options by changing the match making.

Classic and Idle are very different games. I don’t think one should encourage participation in the other.
I’d say clan wars is a Classic event, but with idle rewards and coins. Same goes for ladders.
Imo the rewards should only be coins because they’re more useful - can use coins to buy anything, can’t exchange Idle items. What would a ladder player want with Idle items? Idle players don’t necessarily like the arenas (QM game) and unlikely to get a high ranking in ladders.

From fizzer's perspective, the 2 games should definitely be as close as possible. This is called cross-sales. And from my perspective, CW is an idle competition (the only one we have). Classic players have ladders and Clan League and whatnot.

If any change, coins should be removed from CW rewards so that those classic elite players are encouraged to leave us CW :P

i had a suggestion a while back that there should be cw coins that allow players to unlock custom colours. i agree that cw rewards should not be idle specific.

@Beep Beep: Don’t forget the part where OP derives Elo from past games and then confirms the results are predictive… by back-testing on the same data

CW is an idle competition (the only one we have). Classic players have ladders and Clan League and whatnot.

No. This is like saying an arm wrestling tournament that has a 1st place prize of free NHL tickets is a hockey tournament. No, it's an arm wrestling tournament with a hockey prize. CW is a classic tournament with Idle rewards, just hoping to generate interest in Idle

If you want a true Clan War-style Idle tournament, UserVoice it. (; But from what we see so far, it would be most likely to end up as some kind of tournament style WZIB event - and then the same problems as CW would still occur, the same people would always play, the same players would always win, those not winning would be unhappy

We all understand that in any competition, there has to 1 and only 1 winner and many non-winners, right? I won't call them outright 'losers' b/c many still get meaningful rewards

krinid is wrong. Clan Wars can have arbitrarily many winners. There's currently over 8 billion people who've won so far every single season of Clan Wars and had Final Ascended Idle before Idle even came to exist.

Instead of complaining fruitlessly about the flaws of Clan Wars, we too can join the billions of Clan Wars winners.

I heard the new CP2077 DLC is really really good. You do not need to plan to be available at one of six specific time slots every day, even, to play it.

Edited 9/27/2023 21:05:24

An alternate universe where FWs are eliminated completely (i.e. you get a message that says sorry try again next slot) yields fewer templates options per day.

Literally no one was suggesting that. I am suggesting 2 scenarios:
1) a clan is hogging the timeslot where they hold more than 2/3 + 1 positions or more than 2nd most numberous clan + 1 or even more than 50% of all number of players -- in this scenario we simply find no match to a few of the players from that clan, so that you end up with a timeslot where everyone can be matched. (if after the removal of players the number of players is odd, we just put 1 player from that clan back to the timeslot)
2) there is no such clan and we have odd number of players - we pick a random player and give them the free win.

I can't believe people are trying to fix a system where the best performance is measured by NUMBER OF WINS.
This is not possible by design.

No, I am trying to fix a system that matches people with 80% win rate with people with 20% win rate against each other. Also, the number of wins is not enough to measure performance, who did you win against is also a factor. Having number of wins for rewards on the other hand can be fine if the matchmaker works well.

Later in the thread, it gets even funnier, when OP slowly seems to realize that rating is a better way to determine who's best. And fire is hot.

This happened in your head only. Later in the thread I started publishing actual results from my test code that I was planning to do from the start.

@Beep Beep: Don’t forget the part where OP derives Elo from past games and then confirms the results are predictive… by back-testing on the same data

This happened in your head. I was measuring the ELO from timeslots 0-5521 and then predicting from 5522 onwards while updating ELO after each game as it would happen in a realistic scenario.

Fun fact: with 1 exception, all the people defending the current system in this thread are from the 33%-66% win rate group, and that one exception is in the 66%-100% win rate group. How funny that the people for who the current system works as a system should work for everyone thing the system is good.

Ok now i don't know how to format text good and give responses to specific posts so I am sorry preemptively.

Regarding your post on winrate - percentage of people that stopped playing:

1. you are correlating these results with good and bad matchmaking but you are listing winrate and people that stopped playing. Do you imply that losing a game = bad matchmaking and winning is the opposite?
2. There could be other factors that contribute to someone’s decision to stop playing. Besides personal reasons it could be game/gameplay updates, community sentiment or idk
3. We dont know how many of these are people that continued played with an alt
4. We are missing the sample size. If some of your categories have significantly fewer players than others, the percentages are skewed
5. Correlation does not mean causation. You indicate a correlation between win rate and player retention. We dont know if matchmaking is the direct cause of these dropout rates
6. Your decision to categorize percentages as you have currently is arbitrary, for all we, viewers, know. You say that players outside the 33-66% wr have higher dropout rates but is that difference significant enough to be attributed to matchmaking

4 and 6 come together in the essence that you provide nothing on the statistical significance of your analysis. The differences in percentages, while somewhat small, might be statistically significant if the sample size is large. If the sample size is small they probably are not. Regardless, we do not know

7. Other potential factors that I now think would influence these percentages could be, assuming the differences in percentages are statistically significant
1. experience: perhaps some in the 80-100% bracket left because the game is too easy. and vice versa players in the 0-20% found the games learning curve too steep. These have nothing to do with matchmaking
2. longevity: How long were these players active before leaving? Retention might correlate more with longevity and novelty if the people that left in the 80-100 bracket played for years, and in the 0-20 bracket played for some weeks

You then list people who stopped and who are still playing, in your arbitrary choice of percentages.

I can agree with you that its evident that:
- 40-80% is the sweet spot of finding the game engaging and below a certain percent i would blame it on steep learning curve, whereas higher percent burnout.

But, here are some notes, and some are repeats from above:

- There is selection bias as this is for people who played more than 10 games
- Just because players with certain win rates played an average number of games doesn't mean their winrate caused them to stop playing or continue playing. correlation is not causation
- averages are skewed by outliers. An outlier is a data point that differs by a wide margin from the mass. Median might be a more accurate measure. What do i mean, if you have in the category 40-60% 11 players, where 10 of them played between 100-150 games and 1 of them played 1500 games, the player with the 1500 games skews the average upwards

You then give us the number of days they had been playing in cw.
- the averages you list for days played are surprisingly close, especially for those who stopped playing. The variance between the lowest and the highest is 105 days, which isn't dramatic over the course of a year, or more. Furthermore, I highly suspect that the 80-100% bracket is mostly players with fewer amount of games, and the whole dataset there is not significant, let alone trustworthy and comparable to the rest. This closeness in average suggests that other factors, not win rate, play a significant role in the decision to stop playing
- Even if players have been part of clan wars for a long duration it tells us nothing about their frequency of play within that duration
- the 198 days is so vastly different than the rest its either what i presume, being influenced a lot of accounts with 10-20 games only, or is actually statistically significant in which case those players found the game too easy or faced a burnout.
- You illustrate that the 40-80 group has the most days, by far, in continuing to play. But equating duration with either satisfaction or the quality of matchmaking is misleading. Could be misleading*. It definitely reflects that players who are moderately successful and invested in the game tend to stay longer
- For your cumulative data, on both who still play and who stopped playing, the difference between the lowest win-rate group and the highest winrate group (341 and 378) is only 37 days, which questions how significant win rates are in determing the player’s engagement in the first place
- again population size in each group, not just percentages