Probability Explained (Salty Style With Weird Examples)


#1

One thing that we encounter a lot at Gems of War is people not understanding why they haven’t received mythic cards. As such, I wrote up a (very weird) explanation on how probability affects our drop rates.

As we do not disclose our drop rates, we will use a fictional example to explain the relationship between drop rates and probability. For example, when you visit a rainbow, there will be the chance to either find a pony or a unicorn. Now, unicorns are much rarer than ponies. As such, there is only a 1 in 100 chance to find a unicorn. If someone has a 1 in 100 chance of finding a unicorn at a rainbow, they will always have a 1 in 100 chance of finding a unicorn at a rainbow. The trap that many unicorn hunters fall into is assuming that their previous visits will change the results of their future visits, but this is not the case as each time a rainbow is visited it is fresh and new! Even if you have previously visited 100 rainbows and only found ponies, there is no guarantee that at the next rainbow you will find a unicorn. Each time you visit a rainbow there is still only a 1 in 100 chance of a unicorn being found. Some lucky unicorn hunters will find several unicorns per 100 rainbows, while others will find none. This is purely down to luck and chance.

As a further example, let’s say you’re flipping a coin. There’s a 1 in 2 chance that it lands heads. However, when you flip the coin again it faces tails up. Next time the chance is still 1 in 2. Tails again. The next chance is still 1 in 2. The past results don’t affect the outcome or chance of the next result. That’s how probability and randomness work.

I hope this has helped you all understand how our drop rates work in regards to probability and randomness. With that said, happy unicorn hunting!

TLDR, unicorns.

unicorn


#2

Thanks for the explanation! Also, first! Love, your friendly stats major. :unicorn:


#3

Is it not required for a release in China? :slight_smile:

You forget about pity timer: and Rainbowmaster is there to give a pony if you didn’t catch anyone after visiting 1,000 rainbows. Rainbowmaster is the best. :wink:

(can we have a Rainbow Giant?)


#4

We are still discussing what needs to be disclosed in a Chinese release. It may not be necessary, and even if it is, won’t be released for quite some time. :slight_smile:

We don’t have a pity timer in Gems of War, which is why some people get very lucky with mythics while others don’t.


#5

I have the vague impression that it was required in Japan in the past if i correctly recall… :thinking:
Not sure about today.


#6

Are you saying than Chinese release is postponed for quite some time?

Why hiding the drop rates? I imagine that this thread already answers this question… :stuck_out_tongue:


#7

The Chinese release isn’t postponed, but releases in new countries take quite a lot of time.

We have never disclosed our drop rates, and this thread does not answer the question at all. I picked 1 in 100 as it was an easy set of number to understand.


#8

I wonder if Mithran or some of the other number geeks are going to come in here and destroy your explanation of probability and results. Going to watch this thread with some popcorn handy.


#9

Since the moment I had all the troops in the game.
I’ve “missed” the last 2 Mythics that were released.

In other words…I was successful with every rainbow. (One rainbow I had to craft.) But since having all the ponies… I’ve been pony deprived.


#10

Salty just gave a couple of examples of the Gambler’s Fallacy. It’s sound, from what I got in a quick check. You can put away the popcorn.


#11

Thank you @Grundulum! I tried to make this explanation as simple as possible.


#12

One problem with that. Apple updated the App Store rules last December.


#13

Bad seed and a cursed salt.

Some pseudorandom number generators are just not random enough when it comes to lottery situations due to bad seeding or salt.


#14

How would you test for that and differentiate it from plain bad luck?


#15

It’s not as black and white as a lot of players seem to believe. If apple requires us to we will, but at this time we aren’t.


#16

There are tests for randomness. It is not that hard to differentiate. :rofl:

I won’t be going in details as to why specifically Salty is somewhat wrong, not entirely wrong just a bit inaccurate. But for most people, it is a matter of personal conviction, so I’ll let is go. This time. :wink:


#17

From what Sirrian and Strange once told me, their is a ‘salt’ created at account creation based on a time stamp.
Everytime (?) you hit the server, there is a ‘seed’ generated based on a time stamp.

Some combination of the permanent ‘salt’ + often generated ‘seed’ is fed into the RNG to generate the ‘random’ numbers. If the same ‘salt + seed’ are used, the RNG generates the same sequence of numbers.

Mythic odds ~ 1 in 10k for glory keys

In order to avoid possible cursed salts, there should be at least 10k^2 (100M) unique seeds into the RNG. If they are just using time stamps that might not be the case, and they might not be using a good RNG algorithm.


#18

Okay but why are there unicorns and ponies but no bunnies?
100% legit Bunicorn:
bun

(I gave up on ever getting mythics from chests again, so I am allowed to be like that, okay?)


#19

Your rabbit is beautiful. I love rabbits!


#20

This would mean that every account has a different “luck”?