Has the "10%" estimate for Golds vs. Purples ever been confirmed?
![[GoT] LuckyNumberHat](https://us.v-cdn.net/6030345/uploads/defaultavatar/nL5515UJH9DYJ.jpg)
[GoT] LuckyNumberHat
✭✭✭✭
in The Bridge
Since starting this game a few years ago, I've always been told through common knowledge that the drop rate for Golds in the guaranteed slot of a 10x Premium pull is 10%. This is also then commonly distributed evenly for the Golds in the pack. Same goes for the Purples. That would imply the following distributions in packs:
Event Pack:
10% - Event Gold
45% - Event Purple A
45% - Event Purple B
Tuesday Pack/Theme Pack
3.3% - Gold A
3.3% - Gold B
3.3% - Gold C
30% - Purple A
30% - Purple B
30% - Purple C
And of course there are the odd duck packs like at the end of Mega with 5 Golds and 8 Purples, but you get the idea. But the question remains, has this estimate ever been substantiated or confirmed? (If you've got an affirmative answer, citation would be appreciated. Otherwise we're still at common wisdom and conjecture.) And moreover, does this seem like it should be part of the "loot box" drop probabilities that are published, given that the packs are specifically advertised as offering odds on featured cards?
Event Pack:
10% - Event Gold
45% - Event Purple A
45% - Event Purple B
Tuesday Pack/Theme Pack
3.3% - Gold A
3.3% - Gold B
3.3% - Gold C
30% - Purple A
30% - Purple B
30% - Purple C
And of course there are the odd duck packs like at the end of Mega with 5 Golds and 8 Purples, but you get the idea. But the question remains, has this estimate ever been substantiated or confirmed? (If you've got an affirmative answer, citation would be appreciated. Otherwise we're still at common wisdom and conjecture.) And moreover, does this seem like it should be part of the "loot box" drop probabilities that are published, given that the packs are specifically advertised as offering odds on featured cards?
1
Comments
And the horror pulls here are mainly from people who don't understand what 10% odds look like. None of the stories have been that terrible to be statistically impossible.
Thank you.
Prob 5* crew (per pull): 1.27%
Prob not 5* (100%-1.27%) = 98.73%
Pulls per pack: 10
Prob no 5* over 10 pulls = 98.73%^10 = 88.0%
Thus prob 5* = 100%-88% = 12% chance of a 5* in a single pack
Given that the EU requires these drop rates to be published I can't imagine DB would lie about them. I agree with AviTrek that some players may not grasp how rare 12% chances are. Basically you are an ~8/1 underdog to get a 5* in a single pack. Similarly there is a 28% of zero 5* in a 10 pack weekly offer (88%^10 = 28%).
I don't think you're missing anything; the 12% answer is the same as if you work it the other way of Probability of 5* on pull 1 OR on pull 2 OR on pull 3 ... your approach is more elegant and easier to compute!
Folks can then extend the math to packs rather than pulls. 100% - (88% ^ NumPacks). It takes 5-6 packs to get to 50/50 odds of at least one gold.
I wish it was this simple, but there are two issues with looking at packs like this.
First, We don’t know how the 1.27% extends to event packs. On normal pulls those are the odds of a gold behold. But on event pack pulls you can get both the featured gold and a non featured behold. Is the 1,27% for one of the two types or for both of them combined?
Second we don’t know how the guarantee works. The results of the tenth card are dependent on the first 9 (and if they are like normal pulls the game may sort so the guarantee pull is first if it happens). While the 1.27% may be the odds of a gold (who knows which type) for the first 9 cards, we don’t know the odds on the 10th card that has to be one of the event cards.
Agreed. My assumption for normal packs (which seems correct but I don't have enough data to confirm), the game does 9(or possibly 10 pulls) if you get a 4* or higher that's what you get. If you've gotten no 4/5* then the game gives you a straight 4* drop.
For event packs I assume a similar algorithm, but that last pull is then guaranteed from the event crew. But we have no indication of the odds between 4*/5* for that event drop. The 6500 Dil pack combined with common wisdom makes me comfortable with the 10% estimate. But unless someone has tracked enough event packs we can't be sure.
Someone had written in the forum (but I don't think cited) an explanation that the 9th/10th pull begoldery is just an artefact of pre-sorting the results so that they stack at the end. (which if true partially untangles the complication, but not completely). Like the number of licks to get to the center of a Tootsie Roll Pop, the world may never know...
From experience the beholds are always sorted to the end on normal premium pulls. If you don’t have a behold sorted to the end then you get a direct drop purple (never gold) at the start. From that you can back into pack odds on normal premium pack pulls.
The event packs and 6 crew packs are completely different since both the featured purple and gold crew can be at the beginning or end of the pack, and non featured beholds appear anywhere in the pack.
There should be an autocorrect feature that adds "on average, with a large enough sample" every time someone uses the percent sign. For example:
"There's a 10% of getting a begold [on average, with a large enough sample]! Why didn't I get one when I bought the $10x10 offer???"
At this point, if it's not 10%, I don't even care, because it's probably like 9.97% instead.
Any sources on that? Was that posted somewhere or done privately by @eXo | Frank?
There is nothing I can add to what is already stipulated as drop rates in the game.
But it obviously did work to some degree because you did notice it
I think the issue with the advertised drop rates is that it’s not clear how they apply across a 10-pull where there are additional factors like guaranteed crew.
That has always been my question - of the featured crew, what are the odds of getting a specific one of them? The published rates seem to apply to the other pulls within the 10-pack.
DB's lawyers have decided that they are complying and Shan will not say anything else unless DB is given instructions from Apple or their lawyers that they have to update their published data.
I don’t feel you’re wrong in any way, but I think the hope was that maybe DB would engage with the community without needing to resort to those lengths.
Apple, for example, might not be interested unless you’re claiming purchases aren’t as advertised, which may require a claim for a refund. On the flip side, DB might view that as a fraudulent claim and ban your account if they discovered the situation. Personally I don’t think I’d want to take it down that route, but I’d still like to see if DB will engage.
I completely agree. Unfortunately it feels like DB feels that the less information they provide, the more money they make.
Excuse me while I go bring up the AND bug thread for Shan to ignore for another week.
I won’t chastise people for trying a less adversarial approach to reach out to DB on these types of issue.
If anything, it’s proof that they at least made reasonable attempts to address the matter with the developer directly.
ANDgate - sure, I feel we’re due another thread on that and I will support the attempt.