Because they are not the values that the game
uses, and because of the way proration works. To explain, let's take Archetype (constant 1.00) as an example. What damage do hits of a combo do to her?
Hit | 1st | 2nd | 3rd | 4th | 5th | 6th | ... | 29th | 30th | 31st | 32nd |
Damage (25%) | 0.90675 | 0.87500 | 0.84375 | 0.81250 | 0.78125 | 0.75000 | ... | 0.03125 | 0.03125 | 0.03125 | 0.03125 |
Damage (100%) | 1.00000 | 0.96875 | 0.93750 | 0.90675 | 0.87500 | 0.84375 | ... | 0.12500 | 0.09375 | 0.06250 | 0.03125 |
Ratio | 0.90675 | 0.90322 | 0.90000 | 0.89655 | 0.89285 | 0.88889 | ... | 0.25000 | 0.33333 | 0.50000 | 1.00000 |
So as you can see, this ratio is not constant throughout a combo, which is why I am opposed to using it as an actual guts value.
Somehow incorporating this into a good effective HP estimate does seem useful, but then I'd have to estimate what the average length of combos are, and what the average effect of move proration on those combos is. ...yeah, I gave up on that, but likely, the "true" effective HP is higher than the numbers that we both have.
I think there is also a difference in what we want with those numbers. If it's just for estimating effective HP, then your estimates are better than mine, because yours do incorporate some extra information. If your goal is to understand the system behind the game better, I think my numbers are better suited for that. With information about the proration system, your numbers can be deduced from that anyway (and with a lot more precision).