Kyber 1 Shrinking?

"Starslayer;c-2391504" wrote:

One chart you’re missing is “gac numbers by league before ‘join’”, during the off week.

I’m not sure what you mean by that? Are you looking for a measure of people who log in?

Rank isn’t meaningful for people who haven’t logged in, and CG assigned a rank to anyone who had logged in with the past…. It was either 6 months or a year or so? When new GAC started.

We used to be able to estimate logins in “old GAC” since under that system, everyone who logged in during the 24 hour join window was assigned a new rank for the month, and we could crowdsource finding the bottom of every division. It held remarkably steady around 400k for 2 years of data collection.

But under the new system, that doesn’t work. In some ways, “people who join GAC” is a better estimate anyway, at least if you want to home in on the engaged player base. We know that a lot of the 24 hour actives never moved out of carbonite under the previous system for example.

If you want longitudinal comparisons, you need to try to find a way to make sure you are comparing apples to apples, not oranges, and that is hard.

FWIW we also see ~1-2k variation week to week in the join numbers - hard to say if that is due to data acquisition or real variation in players, but that kind of movement is in the noise.

And I was actually kind of shocked that there are no significant differences in join numbers between 5v5 and 3v3 months - anecdotal evidence is that no one likes 3v3, but everyone is still joining for their crystal I guess!

"Drathuk916;c-2391565" wrote:

"harvestmouse;c-2391552" wrote:
So in relegation out of you get squished for sure.

"Drathuk916;c-2391433" wrote:

"PeachyPeachSWGOH;c-2391420" wrote:

"Drathuk916;c-2391400" wrote:

"MaruMaru;c-2391360" wrote:

"Drathuk916;c-2391302" wrote:

"MaruMaru;c-2391262" wrote:
If this info is true:
"Iirc around 16k people started in kyber 1 in season 1. It’s less than 6k now." by @Drathuk916
Sorry but you can't explain away the change in sizes simply by quitters.

Not entirely clear what you’re taking exception too. First season had around 343k according to swgoh.gg. I forget the poster before Scott over there but you can find it yourself. So around 34kish fewer accounts are signing up each season. Kyber league has shrunk from 88k to under 50k. So yes obviously not all accounts that have quit were in kyber league. However I certainly believe more than half of the accounts that no longer signup were in kyber league. This still leaves a sizable amount of accounts to be demoted through inactive play but still signing up and to much much lesser extent poor players being pushed out of kyber.

I agree the squish has had an intended effect of pushing people out of kyber or more correctly making it more difficult for players from aurodium to play well enough to replace those playing badly enough in kyber who earn a demotion. But I don’t think it is larger than the effect of quitting players is having.

I didn't try to address you directly, but rather commented assuming your info is true. I assumed flat distribution of quitters from all leagues. You might be right that quitting top heavy might be more prominent. Still the end result is it has shrunk to %38 of it's initial size which is a pretty drastic change.

Gotcha. I was trying to clarify my opinion is all.

And to @PeachyPeachSWGOH yes, get to benchmarks that can be moved every six months or so and join but never participate get a better baseline reward. It’s no different than getting a meta team and ignoring squad arena in the past. Well sure it’s giving out more crystals but everyone is already getting more than they use to.

Under this change an 8 million account that joins but doesn’t really participate gets a total of 8960 crystals in 28 days or 320 a day average. Yeah that seems like a lot but that’s still around half of what a k1 active participant gets. You can change the reward structure too if you like. The actual crystal reward doesn’t matter to me. What matters is making it easier for younger/smaller accounts to plan a path forward to a reasonable daily income while lessening their frustration with severely unbalanced matches due to lack of active participation. I mean there is an 11 million gp account in carbonite 5 for deity’s sake. How is that fun?

Well, I don't know about you, if it's between the amount of work one has to do in order to stay in K1, or half of the reward for doing practically nothing? I'd pick the latter.

You would need to steepen the reward ladder by a lot in order to incentivize people to strive for higher divisions. Something tells me CG will not drastically increase K1 payouts so it's the lower Kyber payouts that must drop for the ladder to be steeper. That would then put a lower cap on all the lower league payouts.

And the real problem of a gp floor is not the rewards, actually. It's that it doesn't solve the problem fundamentally, because instead of 11 mil accounts in Carbonite 5, you would just have 7.9 mil ones in Carbonite 5, and the 11 mil ones are just going to hang around Kyber 5. They would still be slacking off, and frustrating other players, except they'd be doing it for much better rewards.

The benchmark would be for each league. 6 million couldn’t drop out of aurodium 4 million chromium 2 million bronzium. It’s the compromise position between the old system and the new system. Your skill ranking can get you into any league or division but your gp prevents you from dropping too far. This is all randomly picked numbers.

So...........I started the new GAC in A3 and I'm currently at 6.6 million. That would leave very little wiggle room with your interpretation. This all ties in ok with those that are competent. However, what about the players that are above average in GP, but not competent?

What happens to them. The players that legitimately played, but fell to Chromium or Bronzium? Under your system it's tough luck.

It's so freaking simple. Deal with those that aren't actively playing and falling and only them. Once that happens see how the land lies.

There's no need to totally go into a GP system; especially for players that didn't suggesting changes that didn't play/understand the old system.

Oh and @Lumiya what actually is your standpoint? Any post that is anti the current GAC, no matter the issue you like...........even mine.

There are issues, yes there are. However, tweaks are needed, not a total overhaul.........in my opinon.

First, you ignored my last sentence of “randomly picked numbers”. But at the same time there was some logic.


The above is from Scott over at swgoh.gg and the rows are in alphabetical order so audorium, bronzium carbonite, chromium, and kyber As you can see the randomly picked number is basically the average for each league gp wise. I also am suggesting a floor not a cap. Your 6.6 million account could still climb to kyber 1 but couldn’t drop below aurodium 5.

The logic is to group accounts that join but don’t attack in division 5 of their appropriate gp level league. If you’re active and in division 5 you should have an easy time climbing out as your opponent is far more likely to not participate.

I don’t see how this suggestion would ever create a scenario where an active account is in a tough luck situation. I also don’t see cg even considering it because it means more crystals and to them I think they believe that automatically means less money. This could very likely be true, but I think player retention both from active and non active joiners would be higher and therefore a wider pool of people who might spend.

I like the posts I agree with, as simple as that. Unlike some here I don't have a problem agreeing with people/liking their comments even if we might disagree on other topics. ?

When I comment I write what I think and also why I think so, so it shouldn't be too difficult to see where I stand on this topic because you have read my comments.

It is a many layered topic but to try to keep it short: The system as it is may be good in theory but not so much in reality. As I stated before on another thread a while ago, a system like this does not work well in a game like this without some kind of filters. I am not saying to go back completely to the old system, but as we can see there needs to be more than just the skill rating, which brings me to another point. The SR gets artificially influenced from the outside after every season. That destroys every self regulation the system might have on top of the other "problems".

But before this post gets too long, I already mentioned all my points with the reasons for why I think so in more detail throughout the various posts that have been made here over the time.

Out if fear my comment disappears if I edit: my comments was supposed to be an answer to @harvestmouse, don't know why it quoted the last post when I actually clicked quote ob harvestmouse's comment. ?

"Ragnarok_COTF;c-2391446" wrote:

"PeachyPeachSWGOH;c-2391442" wrote:

"Ragnarok_COTF;c-2391440" wrote:

"PeachyPeachSWGOH;c-2391418" wrote:

"Ragnarok_COTF;c-2391371" wrote:

"PeachyPeachSWGOH;c-2391187" wrote:
The goal to keep division/league sizes constant was set by CG, not some forumers, though:

https://forums.galaxy-of-heroes.starwars.ea.com/discussion/251844/developer-q-a-grand-arena-championship-changes-12-02/p1

If they are still trying to keep up with that goal, the conclusion would be that the overall GAC population has been shrinking.

Those aren't sizes by count, they're distributions. And as many others have already mentioned, Kyber (K1, in particular) was over-saturated with the initial GP rankings. So of course the head count in divisions will need to change to achieve the desired distribution.

Aren't the distributions percentages of head counts? What's the difference between "sizes by count" and "distributions?"

As for Kyber being over-saturated initially, I have always assumed that they picked the initial GP thresholds so that the population would be approximately evenly spread across the leagues.

If my goal is to keep 10% of X in a group, then the size of the group will be tied to X and won't necessarily be constant. Also, if my initial state is way out of alignment, I will have to adjust the size of the group to get to my desired %. I could do this all at once, but the shock might be too great. So maybe I take my time and spread it out over time.

I don't think we disagree, then. The Q&A I quoted - "evenly distributed", and "10/25/30/25/10" - is obviously talking about percentages. Nobody is talking about fixed numbers, I don't think.

You thought wrong :P



Maybe you meant evenly distributed. But strictly as written, your comment contradicted the quote you were sharing, which I found confusing.

I see where the confusion came from. Sure, I should have said "keep the league/division sizes constant in relation to each other". We have been always talking about ratios and relative sizes.

"Starslayer;c-2391581" wrote:

"Taliana;c-2391524" wrote:

"Starslayer;c-2391504" wrote:

One chart you’re missing is “gac numbers by league before ‘join’”, during the off week.

In some ways, “people who join GAC” is a better estimate anyway, at least if you want to home in on the engaged player base.

Nah, i was trying to find a reason to the kyber shrinking. I wonder if there is a shrinkage in Kyber because we have a shrinkage of engagement in Kyber (which could be a result of datacrons) and not a voluntary shrinkage. Inactive players still receive daily crystal, so if there is a simple 20% per league objective, it might consider the entire population and not only the active population.

Nope, peoples skill ratings are dropping even though they have winning records, because of squish. It is 100% a manufactured situation.

If you follow it from month 1, it is actually the opposite problem - they started out dividing up by GP evenly into leagues. Then the top players mostly kept playing GAC, but the middle dropped out. So kyber was too big as a percentage of active GAC.

So then they introduced squish to presumably force kyber back down to 20% of active players. Which makes some amount of sense, but because they didn’t communicate anything about it, it is perceived as a punishment by the active players.

This is exacerbated by whatever they are doing to kyber 1, which….. makes no sense to anyone. We kind of assumed it would stop at 10% of kyber and kyber at 20%, but it didn’t.

"Starslayer;c-2391605" wrote:

"Taliana;c-2391603" wrote:

So then they introduced squish to presumably force kyber back down to 20% of active players.

That's where I disagree. It might be 20% of players, not active players. Hence the need for a "join or not join" data. If it's 20% of players, the shrink that we can see in your graph could be a drop of engagement by Kyber players.

If you don’t have any data to the contrary trust the data we have collected.

The most telling point is looking at the skill rating of actual players, which has gone down even though they have winning records. Mhann of Mhanndalorian Bot has data for hundreds of players, including his own.

The players currently filling kyber2 are the exact same players who were fighting it out in div 1 before they changed the size on us. The kyber1 players aren’t gone, we are all just in kyber2. And the bottom 5k old kyber2 players are in kyber3. And the old kyber5 players are in aurodium.

I really appreciate all the discussion. I think understand why it is happening, I'm just curious if it's intentional or just lack of understanding of how the skill rating squish is influencing division sizes.

"Taliana;c-2391607" wrote:

"Starslayer;c-2391605" wrote:

"Taliana;c-2391603" wrote:

So then they introduced squish to presumably force kyber back down to 20% of active players.

That's where I disagree. It might be 20% of players, not active players. Hence the need for a "join or not join" data. If it's 20% of players, the shrink that we can see in your graph could be a drop of engagement by Kyber players.

If you don’t have any data to the contrary trust the data we have collected.

The most telling point is looking at the skill rating of actual players, which has gone down even though they have winning records. Mhann of Mhanndalorian Bot has data for hundreds of players, including his own.

The players currently filling kyber2 are the exact same players who were fighting it out in div 1 before they changed the size on us. The kyber1 players aren’t gone, we are all just in kyber2. And the bottom 5k old kyber2 players are in kyber3. And the old kyber5 players are in aurodium.

Exactly and it creates a huge domino effect. Kyber players squished/falling down which in turn pushes everyone else also down.
Total control of crystal flow and less and less crystals going to the players.

The squish is to help CG control overall crystal income in its entirety (i.e. player base).

"MaruMaru;c-2391633" wrote:


This "drop of engagement=drop of league size" doesn't make the least sense to me. In a system that's designed to keep ratios(not numbers) constant, drop of engagement of some would automatically get replaced by comer-uppers.

So the last time I got relegated out of Kyber. I was then squished points to keep me out of A1 and put me into A2. So........ I would then need to promote myself to get promoted back to K5, instead of already being in a position to be promoted.

Maybe this leads to less coming up than going down?