Question on the book of “Revenue management and Pricing Analytics”, page 6, the use of marginal analysis

Recently, I sanction purchased the dimensions "Revenue Management and Pricing Analytics" written by Guillermo Gallego and Huseyin Topaloglu.
I sanction a scrutiny encircling the use of final segregation mentioned on page 6.
By using final segregation, it parallels the anticipated final fruits of generous retribution to the expected final fruits of discounted retribution. And it uses this similarity to declare whether a concourse should sanction the supplicate for discounted retribution or refuse it.
However, I sanction ground this inference a paltry bound, accordingly the announcement in the dimensions involvedly assumes that the y units of competency helpful to generous retribution can disagree. That is to say, for each potential appreciate of y, from 0 to c, we parallel the anticipated final fruits of generous retribution to the expected final fruits of discounted retribution. So, can you construct this involved boldness patent?
Also, can you defend the use of final segregation, if the appreciate of y is unwandering? That is, when the appreciate of y is unwandering, can you defend that the final segregation is an misapply mode to trace an indicator or a criteria that can state whether to sanction or refuse the supplicate for discounted retribution? When the appreciate of y is unwandering, the final segregation singly considers the yth fruits. But how encircling the (y-1)th, the (y-2)th, the (y-n)th fruits?
Thanks!