OMFG Taylor
What testing was done to confirm that a +X luck bonus = a +X% increase in the amount of gold dropped?
Glad you asked!
= Overview =
My hypothesis was that each point in luck increases the amount of gold a monster drops by 1%.
= Method =
I would pick a monster at
CL X.x and suppress myself to
CL X.x. I would only kill that type of monster at
CL X.x for all of my data.
I would first gather data for a control group using an avatar with zero luck. I would get the minimum and maximum gold drop amounts I received in this group and increase each number by my luck boost of 26% or
1.26. The resulting numbers would be my predictions for the minimum and maximum gold drop amounts in the experimental group. I would round all of these results since we never receive fractions of gold in the game.
Next, I would conclude by gathering data for an experimental group using an avatar with a
+26 boost to his luck
(RR4 Divinity, RR4 Coyote Spirit, and Fortune's Favor).
Finally, I would compare my predicted min and max numbers against the actual min and max numbers in the experimental group.
= Execution =
First I picked
CL 3.6 Taiko Drums. I made sure my avatar had zero Ghi and zero bonuses to his luck. I suppressed to
CL 3.6 and
only killed
CL 3.6 Taiko Drums to record the amounts of 50 gold drops for the control group.
The min and max gold drops in this control group were 11 and 22. Multiplying both by 1.26 and rounding I predicted the min and max gold drops for my experimental group to be:
Min:
round(11 * 1.26) = 14
Max:
round(22 * 1.26) = 28
I then gave myself a bonus of 26 Luck
(RR4 Divinity, RR4 Coyote Spirit, and Fortune's Favor), suppressed to CL 3.6 and proceeded to kill only CL 3.6 Taiko Drums to record 50 gold drops for my experimental group.
The result was 14 gold as the min drop and 28 gold as the max drop. Data from drops is below, with control group on the left and experimental group on the right, sorted in ascending order:
11 ___________ 14
11 ___________ 14
11 ___________ 14
11 ___________ 16
11 ___________ 16
11 ___________ 16
11 ___________ 16
11 ___________ 16
11 ___________ 16
12 ___________ 16
12 ___________ 17
12 ___________ 17
12 ___________ 18
12 ___________ 18
13 ___________ 18
13 ___________ 18
14 ___________ 18
14 ___________ 19
14 ___________ 19
15 ___________ 21
15 ___________ 21
15 ___________ 21
16 ___________ 22
16 ___________ 22
16 ___________ 22
16 ___________ 22
16 ___________ 22
16 ___________ 23
17 ___________ 23
17 ___________ 23
18 ___________ 23
18 ___________ 23
18 ___________ 23
18 ___________ 23
18 ___________ 24
19 ___________ 24
19 ___________ 24
19 ___________ 24
19 ___________ 24
20 ___________ 26
21 ___________ 26
21 ___________ 27
21 ___________ 27
22 ___________ 27
22 ___________ 27
22 ___________ 27
22 ___________ 27
22 ___________ 27
22 ___________ 28
22 ___________ 28
I wanted to verify this so I moved on to CL 4.9 Buzz Saws. Repeating the above steps for collecting control group data I received a min gold drop amount of 12 and a max gold drop amount of 25.
Min:
round(12 * 1.26) = 15
Max:
round(25 * 1.26) = 32
I repeated the above steps and gathered the data for the experimental group. The min and max gold drop amounts for this group were 15 and 32 gold respectively. Data from drops is below, with control group on the left and experimental group on the right, sorted in ascending order:
12 ___________ 15
12 ___________ 16
12 ___________ 16
12 ___________ 17
12 ___________ 17
13 ___________ 17
14 ___________ 18
14 ___________ 18
14 ___________ 19
14 ___________ 19
14 ___________ 19
14 ___________ 19
14 ___________ 19
15 ___________ 19
16 ___________ 19
16 ___________ 19
16 ___________ 21
16 ___________ 21
17 ___________ 21
17 ___________ 22
18 ___________ 22
18 ___________ 23
18 ___________ 23
18 ___________ 24
19 ___________ 24
19 ___________ 24
19 ___________ 24
20 ___________ 24
20 ___________ 24
20 ___________ 24
20 ___________ 26
20 ___________ 26
21 ___________ 26
21 ___________ 26
21 ___________ 26
22 ___________ 26
22 ___________ 27
22 ___________ 27
22 ___________ 27
22 ___________ 28
22 ___________ 28
23 ___________ 29
23 ___________ 29
23 ___________ 29
23 ___________ 29
24 ___________ 29
24 ___________ 32
25 ___________ 32
25 ___________ 32
25 ___________ 32
= F.A.Q. =
Q: Why only compare the Min and Max values?
A: I compared only the min and max values because random distribution necessitates a much larger data set in order to draw an accurate comparison between the amounts of gold being dropped for the control and experimental groups.
Q: Why only 50 data points?
A: Since I was only interested in the end points of the gold drop range I needed comparatively fewer data points to find it. I chose 50 because I knew the gold drop amount ranges would be small.
Q: Is this evidence conclusive?
A: I believe so. We can repeat the above experiment on other monsters and gathering more data points, however all predicted min and max amounts are fully expected to be within +/- 1 gold of the actual amounts. These small variations can be attributed to the rounding method used in the game scripts.
Questions or comments? Please post! : D