By way of comparison, I wrote a really quick 6-sided dice sim here that tallies up the distributions for a given set of rolls:
Compare and contrast this with the results you are observing. (I make no claim either way about the RNG at play here.)
5 Likes
Can you do one with âloaded diceâ as well.
We know chaos orbs are weighted.
Why would deeds be any different?
4 Likes
There are a million different ways to tamper with the results of a fair die roll, so Iâm afraid that wouldnât be particularly illuminating.
1 Like
I think that the issue here is that players have just taken a MASSIVE hit to their confidence regarding the accuracy of information provided which has damaged trust.
Combine that with the perceived value of purple deeds compared to basically all other deeds and there starts to be a perceived issue with the low level of purple deeds being received
Add to that only purple deeds have been offered in flash sales - which further adds to the players perception that things are not ârandomâ and we are not being provided factual information (Not saying that flash offers are random, but the fact they have been directed at the âmost valuableâ deed is an attempt to monetize that resource requirement raises the question about all resources being provided randomly)
Purple deeds outcome could of course been randomly achieved (reason: maths/statistical variation allows for any single outcome to exist in the set of possible outcomes), but the above factors cause it to be questioned and the trust issue presents a significant roadblock to acceptance of the explanation
What I am not sure of is how we move things forward from here in a constructive manner, nor how we start to rebuild the trust
3 Likes
Letâs assume no mistake has been made (as per treasure/troops with chaos chards)âŚand that @Saltypatra is telling the truth. Letâs then assume that the âdeed colourâ is rolled first, and the rarity of the task is rolled second.
My question is whether the rarity of the Deed Task is different for different colour deeds?
Or⌠We can assume the data that a player bothered to compile is more accurate than how a dev thinks it works.
Ala Chaos Portals. 
5 Likes
I disagree with your premise since its based on assumption and random events are random
(Totally acknowledge that trust in random is at an all time low currently which does have a factual basis)
HOWEVER
this might point to a specific game design decision that having random colors at random rarity may have been a poor choice
(Perhaps similar to the initial relase of Dungeons which provided a random chance of getting diamonds - which was later fixed to be a guarenteed (lower) reward)
Instead what if there was just a rotation of deeds based on color/rarity that cycle through (like the soulforge mythics)
I know you dont get the player impluse to HAVE to check in each day from FOMO in case today is day that a Purple Mythic Deed will be on the adventure board
But what you exchange for that would be a clear and certain planned path (and timeframe) for players to complete kingdom upgrades as well as certainity for players about what a flash offer purchase would provide in terms of speed boost
(e.g. 24 months for a player to complete the entire kingdom upgrade deed collection, 12 months for a player who buys all the flash offers to complete)
Sample sizes in this system are too small for this data set to prove anything for a long time. The devs have stated that all deed task colors are equally likely; even if theyâre wrong (chaos portals), we wonât be able to confidently reject that claim until itâs past the point of mattering.
Personally, Iâd rather believe that the rates are fair, and focus on more important things like complaining about delves. Maybe in 4.7 deeds will be overhauled enough that we wonât have to worry anymore.
Flash offers being exclusively purple looks bad. But even if purple deeds had been the most common up to this point, weâd still keep getting those purple flash offers, because everyone knows theyâre the most valuable color. I donât think we can infer anything about drop rates from the flash offers.
2 Likes
It could be that the Devs think theyâre being nice by offering the Deeds that have been least frequent as a means of catching up?
Itâs not the same. Deed pulls are still well within reasonable expectations, Chaos Portal pulls were way out of bounds. This really requires a lot of math to understand in detail. Simplifying things a lot, just imagine you go to a restaurant and order a glass of water. For Deeds, the waiter brought you a whole bottle, which is somewhat unusual but still okay. For Chaos Portals, the waiter pointed at the ocean and said âitâs all yoursâ.
2 Likes
I sea what you did there.
Day 83 since first Deeds in AB:
13/102 Imperial Deeds (652 days needed at this rate)
18/146 Blue Deeds (674 days needed at this rate)
32/146 Green Deeds (378 days needed at this rate)
31/176 Red Deeds (472 days needed at this rate)
28/176 Yellow Deeds (522 days needed at this rate)
8/176 Purple Deeds (1,826 days needed at this rate) (16/176, 913 days with ~17USD Flash Offers)
20/176 Brown Deeds (731 days needed at this rate)
6 Likes
Out of curiosity, I did 200 rolls with the 137 value of total deeds received thus far: not once did a single-digit value appear.
Maybe thatâs telling, maybe itâs not. If the believers want to try e.g. 1000 rolls and see how many times they get 8 or less in one of the six slots, then divide that by 6 for the likelihood of that falling on the Magic deeds, and thatâs the multiverse probability that you need to accept if you believe that this state of affairs is ârandomâ.
3 Likes
Out of curiosity, I did 200 rolls with the 137 value of total deeds received thus far: not once did a single-digit value appear.
And this is a classic example of the problem with statistics and probability - Itâs complicated and often misunderstood
A one in a million event does not occur once in every million trials
It is easy to have different example outcome(s) and base an âargumentâ on them â but unfortunately this does not provide any actionable evidence
Statistics talks in terms of Confidence Intervals and Standard Deviations for a reason. It does not use single examples (or even 100âs or 1000âs of examples).
The point I am making is not that the system in place is truly random, it is that this is not the way to show that the system is flawed
For more details, see the work @Mithran did to provide evidence of a problem with drop rate distributions and you will see it is not trivial to produce
3 Likes
I have never trusted Gems of War RNG ever since Dragon Soulâs 25% chance to respawn 6+ times in a row often to quite a lot of people. Thatâs not exactly something that should be âcommonâ.
Thereâs something about their programming that causes RNG things to be very clumpy (like their original gem spawners that spawned gems all near each other way back in the day).
5 Likes
I upgraded Zaejin for Amanithrax plus Goblin team help. +1 Health ainât bad. Hopefully +1 Magic for me in another 83 days. 
Gem spawners arenât completely random, there is a density variable that they can adjust. Doesnât require a patch, at least on PC/Mobile. They cranked it up too âhighâ when they were experimenting, so all gem spawns were clumped together. Thatâs when it got noticed that they can and do this.
Yes, gem spawning is one place where we know the RNG is just an input. They have the âcombo breakerâ and lots of other massaging they do there. So itâs ârandom-likeâ, but theyâve tried to attach some undocumented and vague guarantees to it.
This is one of my favourite things in the forum
2 Likes