Introduction
As most of us know, the Amanithrax Delve (NOT just the Event) was bugged, with the last room’s treasure increment parameters mistakenly set to zero. The developers are now sending out compensation. I tracked every reward I received across the weekend and calculated the actual amount of treasure I missed. Comparing this to the compensation message I just received makes for poor reading. I’ve also seen what guild-mates received, and we can make some tentative conclusions.
Preliminary Conclusions
The compensation is significantly lower than it should be. There are a number of candidates for what the devs may have missed in their calculations:
- Daily Delves played in Amanithrax (always a smart thing to do before starting a Faction launch Event).
- Treasure awarded simply from completing the final room (which is multiplied by the full, final chest multiplier).
- Applying the missing 0.25 multiplier to the rewards that were received.
- Reward differences depending on how many rooms were played.
- Reward differences depending on which rooms were available and played (ie: the actual end-of-Delve parameters).
- “Lucky” loot rolls.
- Possible mismatches between what the devs think the treasure tables are and what is actually implemented.
At this stage, I believe 1 and 2 are certain and significant; 4 and 5 are certain but much less significant; 6 is unlikely; 3 and 7 are likely and significant.
Based on my own actual drops, my observations and calculations, the compensation I was sent is less than half of what I would have received without the bug (and more like a tenth for Gold and Souls).
Calculating the Shortfall
A lot rests on the correctness of my calculation, so it’s important I explain it. Let’s start with Final Room Treasure.
Final Room Treasure is simply 500 Gold, 30 Souls, 6 Glory, multiplied by the final Room Treasure Multiplier (Delve Level has no effect). Since the latter was 0.25 short, we multiply each of these quantities by 0.25. So that’s 125 Gold, 7.5 Souls, 1.5 Glory per completed Amanithrax Delve (whether within the Event or not). (No rounding because the rewards given were already rounded down.)
Delve Completion Treasure Shortfall
Assume Delve Completion Rewards are calculated as follows:
- Roll a treasure type and a random number for each final Chest Level.
- Multiply each by the final Room Treasure Multiplier and round down.
Even though each Chest is a separate “roll”, there’s no way to figure out what those random numbers were, so we have to treat all the treasure from a single Delve as a whole. Effectively, we have to compensate for two things: a) There was one item of treasure less than there should have been in step 1. b) the Rewards received were multiplied by a number 0.25 lower than it should have been in step 2. The latter is easy to calculate, but the former is impossible; instead, we choose to upscale the received rewards by a percentage for the missing Chest.
In summary, we take the observed rewards, divide them by the multiplier actually used, then multiply by the correct multiplier (which is 0.25 higher). Then we increase that by the ratio of the observed Chest level and a chest level 1 higher. Finally, we subtract the received rewards.
Mathematically:
T* = ( T / m * (m + 0.25) ) * (L + 1) / L
and compensation is Tc = T* - T
where
T* = corrected treasure amount
T = treasure actually received (observed)
m = Room Treasure Multiplier (observed)
L = final Chest Level (observed)
Ideally, we just want a multiplier, A, to apply to the observed rewards:
Tc = A * T
So here’s what we end up with:
A = Tc / T = (T* - T) / T
= [ ( T / m * (m + 0.25) ) * (L + 1) / L - T ] / T
= ( 1 / m * (m + 0.25) ) * (L + 1) / L - 1
= ( (m + 0.25) * (L + 1) ) / ( m * L ) - 1
= ( m * L + m + 0.25 * (L + 1) - m*L ) / (m*L)
= ( m + 0.25 * (L+1) ) / (m*L)
My averages were m~3.69, L~7.21, yielding A~0.216
The Numbers
I’ve put all my information into Google Sheets – should be viewable for everyone (I think anyone can make a copy to play with):
Delve by delve data appears on the ‘Full’ tab; the final calculation of missing amount is on the default ‘Summary’ tab, which I’ll explain further.
There’s three ways to calculate the missing Delve Completion rewards (spreadsheet shows all three):
i) calculate the average value for A (0.216) using just the average multiplier (3.69) and chest level (7.21), then apply to the total loot;
ii) calculate A for each Delve, determine its average (0.219), then apply to the total loot;
iii) do the whole calculation for each Delve, then sum.
Finally, add in the missing Final Room Treasure.
Here’s the overall results using the third, most accurate, calculation:
Final Conclusions
Based on these calculations and what my Guildies have said, it looks like the compensation has been calculated only based on Delves completed within the Event, and the numbers used don’t match the reality of my observed rewards at all.
Interestingly, even pushing the average Chest Level up to 9 and the multiplier up to 15 (an impossible number) doesn’t come close to matching what I observed.
Here’s my conclusions:
- The compensation issued awards the following per Delve completed in the Event: 10 Chaos Shards, 250 Gold, 25 Souls, 5 Glory, 0.5 L Ingots, 0.25 M Ingots.
- These numbers are significantly lower than what would result from the rewards I actually received.
- EITHER my understanding of how Delve Completion rewards are calculated is wrong OR my calculations are wrong OR my observed rewards were much higher than average OR Sirrian missed something in his calculations.
- No compensation has been issued for missing Room Treasure for the final room.
- No compensation has been issued for Delves completed in Amanithrax but outside the Event.
- The compensation issued appears to ignore how many rooms were actually done, however even using the maximum values of the parameters falls well short of my observations.
In short, if my calculations are correct, a very simplified formula has been used, significant situations have been ignored, and the parameters in the formula are, unfortunately, just plain wrong.
Next Steps
I would be very interested to know if there’s any flaws in my methodology. If it’s correct, then additional compensation seems to be in order.