Although CouponPages did not answer directly my question, his reply is an excellent edification on the working mechanism of an SSD.
While a curve alluded to by me in my previous post is being obtained, a cell may accumulate a few write cycles with the number of the cycles depending on the number of the data points collected.
My answer was from memory, not research. There is no way to plot speed changes based upon the percentage of available space, because there is no decrease in speed due to the lack of space itself.
The decrease in speed is because of the wear level on the cells. The older an SSD is, the more cells have exhausted their lifespan. As a drive fills up, those cells that have had the same data on them since day one may in fact be on their 1st cycle, but since the last x% have been re-used over and over, many of those cells are close to their end of life.
The best way to demonstrate it would be to say that if you took a brand new drive out of the box... installed the O/S then benchmarked it with 10% used, then arbitrarily filled it with another 75% of dummy data, the speed would likely be identical. But, if you leave the drive at 85% capacity, the remaining 15% will be the only part of the drive that gets constant use, instead of spreading the write cycles to fresh cells.
As such the cells in that 15% would become the slower ones... one cool thing, if you then removed the 75% dummy data, the drive would be fast again. Even more interesting, put the 75% dummy data back... it's still fast because some of the remaining 15% are in fact cells only on their second write... but in time, if you keep this 85% capacity, those 15% cells would wear down too.
That's why I suggested a periodic backup and restore as a short term remedy. It's like shuffling the deck.
