Let's Talk Token Dilution

MaxxPowerz

Now this is something that has been bothering me for awhile - I feel like it is taking longer to max cover new characters.

Maybe you guys can correct me where I'm wrong:

When spending a latest legend token you have an increased odd of pulling a new character (*12 if I'm not mistaken). However with each new character that is added the odds of pulling any specific character overall goes down.

So while this multiplier for new characters might sound good, it is becoming harder and harder to actually cover these new characters that are constantly added to the game?

I feel like I am making more progress trying to cover my 5* characters, which seems a little backwards. Should something be done to address this phenomenon, or am I just confirmation biasing it up?

Player1575

They recently updated the multiplier from 3x to 4x on the latest 12 4*s, so they're updating that as time goes by to give them about the same pull rate as when they stayed doing it this way. Classics on the other hand, are definitely getting harder to pull.

I haven't pulled a single cover for some of my champed 4*s since 2017.

Dormammu

It's a valid question you ask. Dilution is certainly a factor for any 4-star cover outside the Latest 12, but I'm curious to know how the Latest 12 works. Is there two 'die rolls' when using a LT? Example:

Die Roll 1 determines which group the token is pulling from
Group A: Three latest 5 stars
Group B: Twelve latest 4-stars
Group C: Everything else

Die Roll 2 then determines which cover you pull from that group's pool

Or, is it just one 'die roll'? Which would mean more dilution among the 4-stars overall, including the latest 12, with every new release.

MaxxPowerz

Good to know, thank you for clearing that up for me. Not too concerned with adding a few champ levels to my champed four stars. I always seem to pull them regardless.

I've noticed lately that im really getting  backed up on my new characters. I have the iso to champ them by I seem to only pull classics  

Orion

The latest 12 4*s always have a combined 50% odds to be drawn and the rest of the 4*s have the other 50%. So the chance of drawing a latest 4* remains constant, while the odds of drawing a vintage one keeps going down as new 4*s are added.

Daredevil217

Orion said:

The latest 12 4*s always have a combined 50% odds to be drawn and the rest of the 4*s have the other 50%. So the chance of drawing a latest 4* remains constant, while the odds of drawing a vintage one keeps going down as new 4*s are added.

This. Max, it seems you’re just having a RNG streak where you’re pulling more Classics than Latests. It should theoretically even out with time. Also, it might be helpful to track your pulls (if you don’t already) because sometimes it may seem more uneven than it is. I know for me sometimes I think “I can’t seen to stop pulling [unchamped character]” but the reality is I just notice those wasted pulls way more than the ones that are useable or Champ levels because they don’t stay in my queue until the last second like the unusable covers do. 

crackninja

However it IS becoming harder, exponentially so, to cover them from tokens if you fail to do so before they slip into the diluted pools ('Classic' 5s and 'vintage' 4s).

It is definitely getting harder to cover classics, but that dilution is becoming less impactful as time passes, not exponentially harder.
Easiest example is when classics went from having 3 to 4 characters, and then from 4 to 5.  Ignore the fact that actual 5* draw rates are higher now than when this was the case since that is a separate factor, so for the purposes of the example, imagine the 5* draw rate was always 15%.
3 characters in classic pool: 5% to get a cover for OML.
4 characters in classic pool: 3.75% to get OML.
5 characters in classic pool: 3% to get OML.
So we can already see that we lost 1.25% just going from 3 to 4 characters, and then "only" 0.75% from 4 to 5.
Presumably this was part of the reason those 5* pull rates were upped from 10% to 15% in the first place, since it was already looking hopeless to cover the original 5*'s if you looked at the draw percentage tables.  They permanently solved that by giving us vague approximations of tier pull rates and no longer throwing those 0.2% odds of pulling a specific 4* at us.

PenniesForEveryone

For any Latest Legendary Token pull:
Odds of getting a 5*: 15%
   Odds of getting any specific one of the 3 latest 5* characters: 5% each
Odds of getting a 4*: 85%
   Odds of getting one of the 12 new 4*: 50% of that 85% or 42.5% per pull
      Odds of getting any specific new 4*: 3.5% each
   Odds of getting a vintage 4*: 50% of that 85%, or 42.5% per pull
      Odds of getting any specific vintage 4*: 0.8% each

Of these odds, the only number that ever changes is the final 0.8% number.  It will always be the 42.5%/(Number of 4* characters -12)....or really -14 if you count Dino/Howard which aren't in tokens.

For any Classic Legendary Token pull:
Odds of getting a 5*: 15%
   Odds of getting any specific one of the 20 latest 5* characters: 0.75% each
Odds of getting a 4*: 85%
   Odds of getting one of the 12 new 4*: 50% of that 85% or 42.5% per pull
      Odds of getting any specific new 4*: 3.5% each
   Odds of getting a vintage 4*: 50% of that 85%, or 42.5% per pull
      Odds of getting any specific vintage 4*: 0.8% each

Of these odds, both the 0.75% 5* odds as well as the 0.8% vintage 4* odds are decreasing over time as the 5* and 4* pools become more diluted.

HoundofShadow

At least we have feeders and covers from progression and placement in PvP and PvE. Not to forget Bonus Hero... It helps a little. I think the developers already have a tipping point before they make changes to the dilution problem. Wait for it. 

Punisher5784

Bonus Heroes was supposed to be their fix to the dilution issue but the odds of getting a BH are low and it doesn't solve the dreaded 6th cover issue. Lately I have a better chance of grabbing the needed covers from random vaults, which is just silly. 

I wouldn't mind spending 25cp on a custom list of characters that I choose to have a higher percentage or a narrow list of only those 12 in a token. 

HoundofShadow

Just re-read it again... I suppose it's a one time custom list? You have to spend an extra 25CP every time you do this.

fmftint

Their experts and metrics have determined it's fine so it's fine

Jrlrma

Heroic and latest keep the new 4x boost, classics rotates boosted odds classic 4s. Simple and easy. 
 You could rotate it (the grouping) every 2 weeks and it would encourage spending since youd have a ton of vets scrambling to get their pulls while their classic toon has their odds boosted. 

Phumade

Why worry about hypothetical math, when there is a real thread with people reporting exactly how many covers they get of a specific 4 that entered and exited the latest pool?

I.e.  How many vultures did you get and are you happy with the volume?

You can all read the responses for your self.

The overwhelming majority of players easily champed vulture (Who was the first 4* to enter and exit the latest pool of 4*.


My vulture easily finished past 300,  so I don't have any concerns about dilution for any character entering 

After 12 more releases, we can ask for more real world reports vs making up hypotheticals.

Phumade

aesthetocyst said:

Why wait 12 more releases? I'm sure we'll have a thread for every character that takes the walk of shame to the vintage vault. Already had a few, though Vulture is the first to be born, live, and die under the current regime.

See you next season for tales of how it went with Mockingbord (and Yondu? haven't looked at how dates are falling...)

I agree with you completely.  if we want to talk about the consequences of token dilution.  The answer is best understood by looking at actual character distributions reports.

Vulture should be used as the baseline (I,e. use your vulture not my vulture) build level because its the first 4* to enter and exit after the changes to vaulting.

How would we incorporate Mockingbird/Yondu levels to describe a dilution factor?
- intutively we know new 4* should be at a lower build because of dilution, but how can we do that mathmatically?


My main point,  is no need to come up and discuss hypothetical theories, when there is a real world data set that is a direct result of the dilution question.

Straycat

MaxxPowerz said:

Now this is something that has been bothering me for awhile - I feel like it is taking longer to max cover new characters.

Maybe you guys can correct me where I'm wrong:

When spending a latest legend token you have an increased odd of pulling a new character (*12 if I'm not mistaken). However with each new character that is added the odds of pulling any specific character overall goes down.

So while this multiplier for new characters might sound good, it is becoming harder and harder to actually cover these new characters that are constantly added to the game?

I feel like I am making more progress trying to cover my 5* characters, which seems a little backwards. Should something be done to address this phenomenon, or am I just confirmation biasing it up?

 If you have 8/12 latest champed, it will take longer to max cover new characters because there are less new characters. I don't know your roster, but I would guess thats where your confirmation bias is coming in. You are trying for a minority of the latest 12, so it is taking longer, while you probably are still pulling 50/50 latest vs vintage