On my machine, instead of putting the Directories right after $MFT as in the Fast Optimize script, I made a 0.5% zone between $MFT and the directories and then dismount the drive (or restart the computer for the C drive). Now instead of one 12.5% reserved zone, there is one 0.5% zone after $MFT.
I have also done this. I pin the $MFT to an Absolute location, and the directories follow after a gap. I believe that as the used $MFT grows, an occasional SlowOptimize [Or Semi-Fast script see below] using that Absolute + gap combination will re-grow the MFT reserved area, by pushing directories back to maintain that gap. (Then reboot after the SlowOptimize).
This makes it practical to keep the Reserved Area quite small; I'm using 100 MB.
Since I don't like all the disk activity during a full SlowOptimize, I'm working out a script for a "Semi-Fast" Optimize. My basic scheme is a gap before $MFT for new files, which gets vacated both in my Fast and my Semi-Fast scripts. After gap - $MFT - gap - dirs, I'm thinking of a zone for recently modified files, that would be FastFill'd in Fast script, and SortByName in Semi-Fast (so that changing files in same dir are together). Then a large zone for the non-recently-modified files: this zone is initially fully sorted by SlowOptimize, but is only Defragment + FastFill in Fast & Semi-Fast. Finally the Space Hogs zone.
The idea is that Semi-Fast would concentrate on optimizing the heavily used or changing stuff, but let the bulk of the files remain untouched. Maybe run this once a week, with a 7-day window for "recently used". (And Fast script run daily; it just defragments, and vacates the new file gap.)
What I'm avoiding is the bulk of the SortByName work done in SlowOptimize. By not forcing every file to be moved, just because a few files have changed. I think I can run for months this way, before there would be any significant gain by doing a full SlowOptimize.