Ok IMC. It seems like your people responsible for math are either incompetent or prefer to just throw around random numbers. This leads to the new DPK system being the most horrible experience ever created.
The current system
The current system is a map-wide DKP with probability
p = 1/DPK
which means for orange 315 weapon materials which have 1000 DPK now the drop chance starts with
p = 0.1 % = 0.001
and after 900 kills reaches
p = 1% = 0.01
for a halfway decent estimate we choose p=0.5% = 0.005 and the binomial distribution (easy to get number examples here http://www.wolframalpha.com/widget/widgetPopup.jsp?p=v&id=226badd3f3410155b258766cca6002ea&title=Binomial%20Probabilities&theme=blue&i0=900&i1=0.002&i2=1&podSelect=)
gives us a
drop chance of DPK before killing 900 mobs = 1 - P_900(X < 1) = 54%.
This means about half of the time you kill 900 monsters without seeing a DPK drop. Which takes hours about 2-3h for most maps because the spawn is super bad. For comparison the same approximation for 600 kills which an average of 1/750 drops gives us a
drop chance of DPK before killing 600 mobs = 18,4%
in other words most likely not going to happen to see something in the first 1-2 hours farming.
It’s horrible. And then someone steals your drop. It’s hell.
And it’s easy to fix!
All that needs to be done is to implement a real Pseudo-RNG system. And always keep in mind that player numbers are low!. Who the bloody hell thought 20k DPK is good for Hunting Ground? Remember, if there is just 1 group grinding there it will never see anything. And that’s how it will be most of the time which is completely frustrating. Ok, the promised fix.
We would like a player friendly system that works even without maps bloated with players (which could have happend if the new system wouldn’t make most quit directly again by the way). For this we define a target % drop change for an item
p = 1/DPK
this means the mean value of expected drop is the DPK which should be fine. To be honest this would already be enough at the moment. But the latest post says that you don’t want us to pray to RNGesus which we have to at the moment because if someone else farming appears it comes to luck who gets it and leads to hate and frustration. So now comes the interesting part to make it player friendly which is an exponential scaling from a number lower than p to a high value for example 10%. Personally I would set the lower bound
b = p/2
as this is basically the same IMC did by doubling the old DPK values at the moment and as upper bound
t = 0.1
and 10% drop should be sufficient to get a drop sooner or later.
Now we need a function that scales those values properly and best for this is the exponential function
current drop chance = c * exp(a * x)
with x being the bound map-wise kill count at the moment. Solved by
c * exp(a * DPK) = t
This function can easily be implemented as it is very cheap to calculate the actual droprate because
c = b
a = ln(t/b) / DPK
Number Example
DPK = 500 (which should be good for orange materials and not frustrating given the spawn rate on maps)
P = 0.002
b = 0.001
t = 0.1
exp = nobody will read that anyways.
But you will get a function that gives you:
0 kill count = 0.1% dropchance
100 kill count = 0.25% dropchance
300 kill count = 1.5% dropchance
500 kill count = 10% dropchance
Should be fair enough to get a drop by luck early on but most likely close to the DPK value. Problem fixed. Thanks!
Little Sidenote
Best would be to remove super rare hour long farm mats all together and use regular drops with a higher number. But guess that’s not wanted. 
End
If you like the idea please also tag the staff and make them read it and maybe forward it to the dev team.
@STAFF_Yuri especially for you as you made the recent Q&A.
