This is How You Fix Dilution

bbigler

I'm writing this in hope that the dev team reads it (as advertised):

The problem: players progress quickly and want to build the 4* characters, but because there are 69 of them (and growing) and most players only get 10 - 20 covers per week, so it takes forever to fully cover any one of them.  This same kind of problem exists with older 5*s.

Failed Solutions: Vaulting = only including the latest 12 in tokens.  This obviously worked for building the newest 4*s, but the older 4*s stopped in their tracks.  Fix #2 Increased Odds for the latest 12 in tokens.  This hybrid approach still didn't work well either.

My Solution: First, combine Latest & Classic tokens into 1 store costing 25 CP.  Second, allow players to choose a subset of characters to pull from for that store (this subset can be changed at any time - the default subset would be the latest).  The subset must contain exactly 8 x 5*s and 20 x 4*s.  This is roughly 1/3 from each tier and the size of the subset would not change as more characters are released.  In addition to this change, make the new 5* store contain only the newest 5*, not 3 x 5*s.  This change would have the following benefits:

• Dilution is no longer a problem, but the pool is large enough to ensure the random nature of the tokens
• Players would be super, super happy to be able to target the characters they want to cover
• As characters get champed, players can decide to either change their character choices or just add champ levels
• This would eliminate hoarding since it would no longer be an advantage
• If players just want the newest 5*, then they can buy tokens from that specific store when the character is released.
  

abmoraz

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

broll

Before the saved covers feature I would have said they'd never have a feature where you can choose like that.  However, we can now save up to 100 covers for an unchamped character so what do I know?

It's a cool idea, and it would certainly give control, I think it would also hurt the game.  People keep coming back (whether they realize it or not) because they get dopamine hits from the slot machine gamboling.  Take that away completely and I think there would be a gradual drop off of engagement that would eventually be a mass exodus if not corrected soon after.  Either because the draw was gone or people achieved their goals too quickly. I've already heard several in my alliance say they are starting to see this just after the saved covers change.  They no longer feel pressure to try and beat the system and that was why they kept playing, trying to beat the house.

Alternative ideas:

Why not just increase the % chance of BH by a significant margin and instead of it being a complete bonus cover you have a like 25% chance to get that character every pull.

Break up the covers into multiple smaller vaults as abmoraz and many before him has said.

My personal favorite, create higher tiers more frequently and as a new tier comes out increase the odds of the previous tier.  This will never fix the 4* tier, they missed the boat on doing it this way for that tier well over a year ago IMO, but let's at least get a cadence to prevent it from continuing to worsen or happen to other tiers.

bbigler

abmoraz said:

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

I originally thought of suggesting it the way you said here, but how do you handle the most recent grouping if it's not an even 6 x 5*s and 15 x 4*s? Will the latest group grow to double size and then get split?  If so, it would be "diluted" most of the time compared to the other groups.   Even having a latest 21 characters group wouldn't fix the problem of some characters being in a diluted group:

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 60 + 5*s #19 - 24
Group 5: 4*s #61 - 69 + 5*s #25 - 26 - the problem, group is too small

Group 4&5: 4*s #46 - 69 + 5*s #19 - 26 - also a problem, group is diluted compared to others

broll

bbigler said:

abmoraz said:

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

I originally thought of suggesting it the way you said here, but how do you handle the most recent grouping if it's not an even 6 x 5*s and 15 x 4*s? Will the latest group grow to double size and then get split?  If so, it would be "diluted" most of the time compared to the other groups.   Even having a latest 21 characters group wouldn't fix the problem of some characters being in a diluted group:

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 60 + 5*s #19 - 24
Group 5: 4*s #61 - 69 + 5*s #25 - 26 - the problem, group is too small

Solution to that is simple. The final group is almost always diluted. 

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 71 + 5*s #19 - 26

When both 4*s and 5*s are double or more the other groups (15 4*s & 6 5*s from your example) then split off a new group. Rinse and repeat ad infinum. 

PS - You keep quoting 69 5*s there are currently 71, Nebula will be 72. 

Jaedenkaal

I can't imagine they'll ever split 4*s like this, if only because 4*s aren't perceived as being equal value, so this would be an easy way to, for example, never draw War Machine ever again. Tools to avoid drawing the characters we don't want have never been something we've been given.

Drlex

Why not just split the 4* into two or three groups. The groups can be random and can switch every so often. Each week you pick the group you want. The tokens, CP, or HP can be used for that group only. Then you could do the same with the classic 5*s. Leave the Latest Legends alone.

bbigler

broll said:

bbigler said:

abmoraz said:

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

I originally thought of suggesting it the way you said here, but how do you handle the most recent grouping if it's not an even 6 x 5*s and 15 x 4*s? Will the latest group grow to double size and then get split?  If so, it would be "diluted" most of the time compared to the other groups.   Even having a latest 21 characters group wouldn't fix the problem of some characters being in a diluted group:

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 60 + 5*s #19 - 24
Group 5: 4*s #61 - 69 + 5*s #25 - 26 - the problem, group is too small

Solution to that is simple. The final group is almost always diluted. 

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 71 + 5*s #19 - 26

When both 4*s and 5*s are double or more the other groups (15 4*s & 6 5*s from your example) then split off a new group. Rinse and repeat ad infinum. 

PS - You keep quoting 69 5*s there are currently 71, Nebula will be 72. 

Currently, there are 68 x 4*s in the Legendary Token pool.  Limited characters, being Devil Dino and Howard don't count because they're not in packs.  Emma Frost was #69, which is why I used that number.  Nebula would be #70.

Also, having the final group grow until double size and then split, being diluted most of the time, I already commented on and said that's a bad idea too, especially since that group would contain the newest and thus least covered characters.  Nevertheless, any kind of grouping, whether pre-defined or user-defined would be better than what we have now. 

Bowgentle

I don't want my 5s tied to 4s.

I draw Latest based on what 5s are in there, I don't care about the 4s.

Unless the current tokens stay in addition to extra types of tokens, I'm not interested.

bbigler

Bowgentle said:

I don't want my 5s tied to 4s.

I draw Latest based on what 5s are in there, I don't care about the 4s.

Unless the current tokens stay in addition to extra types of tokens, I'm not interested.

Not everyone thinks that way.  I think a grouping would add more strategy to the game, making you decide which pool you want to go after, weighing the pros and cons of the best available 4*s or 5*s in it. 

Currently, it's extremely difficult to fully cover Classic 5*s.

Bowgentle

bbigler said:

Bowgentle said:

I don't want my 5s tied to 4s.

I draw Latest based on what 5s are in there, I don't care about the 4s.

Unless the current tokens stay in addition to extra types of tokens, I'm not interested.

Not everyone thinks that way.  I think a grouping would add more strategy to the game, making you decide which pool you want to go after, weighing the pros and cons of the best available 4*s or 5*s in it. 

Currently, it's extremely difficult to fully cover Classic 5*s.

You can ignore 90% of classic 5s.
Feeders will eventually finish Thanos and Thor, who basically are the only classic 5s you need.

broll

bbigler said:

broll said:

bbigler said:

abmoraz said:

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

I originally thought of suggesting it the way you said here, but how do you handle the most recent grouping if it's not an even 6 x 5*s and 15 x 4*s? Will the latest group grow to double size and then get split?  If so, it would be "diluted" most of the time compared to the other groups.   Even having a latest 21 characters group wouldn't fix the problem of some characters being in a diluted group:

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 60 + 5*s #19 - 24
Group 5: 4*s #61 - 69 + 5*s #25 - 26 - the problem, group is too small

Solution to that is simple. The final group is almost always diluted. 

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 71 + 5*s #19 - 26

When both 4*s and 5*s are double or more the other groups (15 4*s & 6 5*s from your example) then split off a new group. Rinse and repeat ad infinum. 

PS - You keep quoting 69 5*s there are currently 71, Nebula will be 72. 

Currently, there are 68 x 4*s in the Legendary Token pool.  Limited characters, being Devil Dino and Howard don't count because they're not in packs.  Emma Frost was #69, which is why I used that number.  Nebula would be #70.

Also, having the final group grow until double size and then split, being diluted most of the time, I already commented on and said that's a bad idea too, especially since that group would contain the newest and thus least covered characters.  Nevertheless, any kind of grouping, whether pre-defined or user-defined would be better than what we have now. 

Good call my bad.  I was just looking at character count and forgot to subtract limited and new characters.

IMO the doubling isn't that bad.  At worst using your numbers there would be 30 4* s in a tier which is far better odds than 68 and growing.  Making the newest harder to get would be what helps drive people to keep trying (or paying) for more pulls which is good for engagement and financial longevity of the game.  We don't know the reason, but the devs walked away from the Latest concept, it's possible it was hurting these factors to some extent.

Another alternative to doubling would be to add each new character to a vault round robin style.  This would also to some extent help with Bow's concern as to stay current you'd have to pull from different groups and not just live in one forever (the newest).  So in your example:

Group 1: 4*s #1 - 18 + 5*s #1 - 7 
Group 2: 4*s #18 - 34 + 5*s #7 - 12
Group 3: 4*s #35 - 51 + 5*s #13 - 18
Group 4: 4*s #53 - 68 + 5*s #19 - 25

Releases going forward would go:
4* release - added to Group 2
4* release - added to Group 3
5* release - added to Group 2
4* release - added to Group 4
4* release - added to Group 1
5* release - added to Group 3

Then once a critical mass reaches they add a new group and reshuffle (potentially even randomly) to create 5 groups.  

bbigler

broll said:

Another alternative to doubling would be to add each new character to a vault round robin style.  This would also to some extent help with Bow's concern as to stay current you'd have to pull from different groups and not just live in one forever (the newest).  So in your example:

Group 1: 4*s #1 - 18 + 5*s #1 - 7 
Group 2: 4*s #18 - 34 + 5*s #7 - 12
Group 3: 4*s #35 - 51 + 5*s #13 - 18
Group 4: 4*s #53 - 68 + 5*s #19 - 25

Releases going forward would go:
4* release - added to Group 2
4* release - added to Group 3
5* release - added to Group 2
4* release - added to Group 4
4* release - added to Group 1
5* release - added to Group 3

Then once a critical mass reaches they add a new group and reshuffle (potentially even randomly) to create 5 groups.  

Hmmm.....not a bad idea but when a group is shuffled, players in the middle of covering those characters will get mad I think.  Honestly, I just want the devs to give us any kind of grouping so that we can pick what group to go after and cover those characters reasonably quickly.  This would also make Classic 5*s obtainable. 

broll

bbigler said:

broll said:

Another alternative to doubling would be to add each new character to a vault round robin style.  This would also to some extent help with Bow's concern as to stay current you'd have to pull from different groups and not just live in one forever (the newest).  So in your example:

Group 1: 4*s #1 - 18 + 5*s #1 - 7 
Group 2: 4*s #18 - 34 + 5*s #7 - 12
Group 3: 4*s #35 - 51 + 5*s #13 - 18
Group 4: 4*s #53 - 68 + 5*s #19 - 25

Releases going forward would go:
4* release - added to Group 2
4* release - added to Group 3
5* release - added to Group 2
4* release - added to Group 4
4* release - added to Group 1
5* release - added to Group 3

Then once a critical mass reaches they add a new group and reshuffle (potentially even randomly) to create 5 groups.  

Hmmm.....not a bad idea but when a group is shuffled, players in the middle of covering those characters will get mad I think.  Honestly, I just want the devs to give us any kind of grouping so that we can pick what group to go after and cover those characters reasonably quickly.  This would also make Classic 5*s obtainable. 

Yeah people might get mad around the shuffle, especially the first few times, but most would get used to it after a few cycles.  Biggest problem would probably be the people that refuse to hoard might have some problems getting what they want because the groups keep shifting.  People who hoard I imagine would just hold onto their hoards until an optimal (or at least good enough) shuffle came out.  It's not all that different than the limited 5* stores on new release, expect spreading the concept out into 4*s and making it a more permanent fixture. 

For people who don't like those vaults, they could keep classics as is and people that don't like the grouping concept can pull from diluted pools at a discounted rate.

abmoraz

bbigler said:

abmoraz said:

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

I originally thought of suggesting it the way you said here, but how do you handle the most recent grouping if it's not an even 6 x 5*s and 15 x 4*s? Will the latest group grow to double size and then get split?  If so, it would be "diluted" most of the time compared to the other groups.   Even having a latest 21 characters group wouldn't fix the problem of some characters being in a diluted group:

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 60 + 5*s #19 - 24
Group 5: 4*s #61 - 69 + 5*s #25 - 26 - the problem, group is too small

Group 4&5: 4*s #46 - 69 + 5*s #19 - 26 - also a problem, group is diluted compared to others

Groups 1 thru 4 are as you said.

Group 5 becomes: 4*s #55 - 69 + 5*s #21 - 26.

So every group has 15x 4* and 6x 5*, but group #4 and group #5 have some overlap in their covers; in this case: 4* #55-60 are in 2 different packs as are 5* #21-24.  If you are targeting one of those covers, you will have a choice of which pack to purchase.

bbigler

abmoraz said:

bbigler said:

abmoraz said:

Don't let the player choose.  Just offer the packs based on when they were released.   Have a "First grouping" pack, which has the first 6x 5* and the first 15x 4*.  Then offer a "second grouping" that has the 7th released 5* thru the 12th released one (and the 16th - 30th 4*), and so on.

keep the latest 3x 5* and 8x 4* in a "latest pack".  This might mean that some of them are doubled up, so be it.  The odds will be slightly different between the 2 packs as will the other rewards.

I originally thought of suggesting it the way you said here, but how do you handle the most recent grouping if it's not an even 6 x 5*s and 15 x 4*s? Will the latest group grow to double size and then get split?  If so, it would be "diluted" most of the time compared to the other groups.   Even having a latest 21 characters group wouldn't fix the problem of some characters being in a diluted group:

Group 1: 4*s #1 - 15 + 5*s #1 - 6
Group 2: 4*s #16 - 30 + 5*s #7 - 12
Group 3: 4*s #31 - 45 + 5*s #13 - 18
Group 4: 4*s #46 - 60 + 5*s #19 - 24
Group 5: 4*s #61 - 69 + 5*s #25 - 26 - the problem, group is too small

Group 4&5: 4*s #46 - 69 + 5*s #19 - 26 - also a problem, group is diluted compared to others

Groups 1 thru 4 are as you said.

Group 5 becomes: 4*s #55 - 69 + 5*s #21 - 26.

So every group has 15x 4* and 6x 5*, but group #4 and group #5 have some overlap in their covers; in this case: 4* #55-60 are in 2 different packs as are 5* #21-24.  If you are targeting one of those covers, you will have a choice of which pack to purchase.

Overlap?  Good idea, why didn't I think of that? 

axmoss

I think they should make a weekly rotating store. 1 latest 5, 2 classic 5s, 12 random 4s that are 12x, the remaining 4s at 1x.

bluewolf

Does this idea allow you to limit your need for roster slots?  Yes.  Are roster slots the #1 driver of spending in game?  Yes.

Sadly, any idea like this will never therefore be implemented.

Perhaps a better/more generous bonus reward structure would be considered.

bbigler

bluewolf said:

Does this idea allow you to limit your need for roster slots?  Yes.  Are roster slots the #1 driver of spending in game?  Yes.

Sadly, any idea like this will never therefore be implemented.

Perhaps a better/more generous bonus reward structure would be considered.

I wouldn't say that it limits your roster slot need because all 4*s and 5*s are required in PVE events, plus we have 4* PVP events too.  Plus, some 4* champ rewards give 5* covers, so that's another reason to roster all 4*s.  So, I don't think this change would affect their bottom line, but it would allow people to champ 4*s and 5*s faster.  The current system to champ 4*s and 5*s takes a very long time. 

bluewolf

What I meant was that the removal of weighted odds for any 4 (when “Latest” odds were removed) means that you have an equal chance of getting any of 70 characters when you get a 4.  Saved covers largely removed the concern of opening tokens where you might waste covers or wait until you had the iso.  The only thing that might stop someone from opening now is having roster space, and therefore, the odds are that you are forced into buying a roster slot if you don’t have self control.  Or at least, feel the need/want to buy that roster slot.

It is true that making people essential is also a reason to roster everyone, but the new dilution also makes it likely that you will see everyone as you build your roster.

My view is that the trade off we received as players for saved covers was that we would feel more roster slot pressure when opening.

I am not going to disagree that the current system is lousy for anyone coming up from 3* tier.  

Farmerbink

In general, there are lots of good thoughts in here.  I don't have an awful lot to add, but support for the concept.  Anything that helps the 3-to-4 transition not take literally years is a plus in my book.