I have a list of random numbers. I sort them, and designate each one as belonging to a tier of some sort (quartiles, percentiles, or even arbitrary cutoff points by proportion) based on the cumulative sum of them in that order. What is this kind of analysis called?
This is not ordinary rank, as I don't care about how many there are in each group nor what their numerical position in the sorted list is.
Example:
Let's say I define four buckets: A = top 16%, B = next 23%, C = next 28%, and D = the rest (33%). I take 20 random numbers from 0 to 100, sort them, and slot them into these buckets according to the running total of them:
Value Cumulative total Bucket Notes
89 89 A
84 173 A 173 > 1014 * 16% = 162.24 ; next bucket
76 249 B
74 323 B
65 388 B
64 452 B 452 > 1014 * (16 + 23)% = 395.46 ; next bucket
63 515 C
57 572 C
56 628 C
54 682 C 682 > 1014 * (16 + 23 + 28)% = 679.38 ; next bucket
54 736 D
47 783 D
47 830 D
45 875 D
45 920 D
41 961 D
26 987 D
13 1000 D
7 1007 D
7 1014 D