You're using reverse logic, but I'll accept your premise as not needing debate for the time being.
The wording does not say a grounded conductor can carry its 90?C ampacity. It says the grounding conductor's ampacity may be determined using 100% of continuous plus non-continuous loads. The grounded conductor is still subject to termination temperature limitations, and correction and adjustment factors, as they apply.
This is where your premise starts to become flawed. You are correct in the deration part, but then you turn around and enter insulation rating back into the picture. Under the same conditions, a 75?C-rated conductor carrying a continuous current which causes it to reach a sustained temperature of 75?C will be the same size as a 90?C-rated conductor (of the same conductor type, e.g. copper). That is, the insulation rating has no bearing on the conductor temperature for a given amount of current and installed under identical operating conditions.
It is quite easy to determine the minimum size conductor for terminal temperature limitation, other factors withstanding. Simply go to the Table and out of the appropriate 75?C column (for 75?C rated terminals) determine the size conductor which has a 75?C ampacity equal to or greater than the actual maximum connected load as calculated under Article 220 (without applying the 125% continuous load buffer). Let's use the OP'er scenario of Posts #1 and #4 as an example. Looking to the Table for a 100A maximum connected load (aka actual load). Thus, the minimum size copper conductor would be a copper-type #3, period. That's it, we're done with this part. A #3 conductor is the minimum size that can be used for this circuit.. But that is under ideal conditions of usage. So, any "selection and coordination" beyond this is for other, non-ideal conditions which must be considered.
I fail to see any benefit of making this a five step process.
The rest is quite simple (as the code is currently written). Simply add in any "125%" buffering to the load value, and divide by the factors for correction and adjustment to determine the minimum conductor size by the resulting ampacity value (note this value is not the circuit ampacity or any other relevant ampacity under Code; it is simply a number to compare with Table values). Going back to the example... we know all of the load to be continuous, So 100A ? 125% + 0 ? 100% = 125A. At this stage, we could look at the Table and see we could use copper conductor types sized at
1/0 60?C-rated,
#1 75?C-rated, or
#2 90?C-rated. All three of the sizes are larger than the #3 minimum we determined above for the terminal temperature limitation.
From there, we also know that if the ambient temperature is higher than 30?C and/or any derating for more than three conductors come into play, that the conductor sizes that we would have to use may even be larger yet, but defintely not smaller. At this point I am simply belittling these aspects to shorten my post. Yes there could be a condition where the ambient is less than 30?C allowing one to use a smaller conductor, or the load is non-continuous. The point I am trying to make is that it is not necessary to draw the process out to five steps just to find the minimum size conductor for terminal temperature limitations, and even then, you still have to include the other size-altering factors before a total correlation can be achieved. It truly is a two-step only process

- Determine minimum size conductor based on maximum connected load current as calculated under Article 220 (i.e. no 125% continuous load buffering).
- Determine minimum size conductor for insulation used by factoring in for continuous loading, ambient temperature correction, and multiple conductor adjustment.
- Step 2 shall not be smaller in size than Step 1.
OOOps! I guess my counting is off this morning. As you can see, it is a three step process
