Threaded Mode | Linear Mode

Pekis · (This post was last modified: 05-24-2019 02:14 PM by Pekis.)

(05-24-2019 12:34 PM)Albert Chan Wrote: Hi, Pekis.

Technically, even if distributed amounts added exactly to total, it is still not equitable.
Rounding operation will slightly favored whoever is lucky.

To force distribution of leftover pennies, I would have added them to B1, B3.
When scaled back to $390, we get this: (max abs err = 0.0048, averaged abs err = 0.0036)

Code:

105.99 → 83.0042, err = +0.0042 72.78 → 56.9964, err = -0.0036 135.36 → 106.0048, err = +0.0048 111.09 → 86.9982, err = -0.0018 72.78 → 56.9964, err = -0.0036

Pennies to B2, B4 is slightly less equitable: (max abs err = 0.0060, averaged abs err = 0.0041)

Code:

105.98 → 82.9964, err = -0.0036 72.79 → 57.0042, err = +0.0042 135.35 → 105.9970, err = -0.0030 111.10 → 87.0060, err = +0.0060 72.78 → 56.9964, err = -0.0036

BTW, why B2 get the penny, but not B5 ? (Both started with the same $57)

Hello, Albert,

you're right, in my example, some are favored (but there is no choice, you'll have to favor some entities in the end !), and perhaps more than their own weight, but I wanted initially to get rid of it in Excel with a simple formula, and without having to choose ...

Thanks for the remark.