That works for players, though I think that making everyone keep a separate piece of paper just for this is just a band-aid solution to a fundamental complexity issue that still exists.

For the DM, however, it's still a whole lot of complexity to track. I know it's hard to test this when you don't have the full system built, but I can basically guarantee that any AOE debuffs using a -5/-10 save progression system will slow the game down a lot the instant they are cast. I mean, think about how much the game already slows down when something basic like a fireball or fear is cast on group of more than a couple creatures. That's a binary pass/fail effect, and rolling the saving throws for every monster and tracking the effects can still take time. Naturally, this depends on how much your playgroup uses debuffs, and how much patience you have, but I think this will make spells a fair bit more complex.

If you're trying to avoid binary effects (which is not a bad thing in principle), I'd recommend using only a "fail by 10" threshold instead of a "fail by 5" threshold; that way, it doesn't come up as much, which reduces how much complexity you have to track. Each spell would have a default condition that it would inflict on a failed save, and it would worsen if the "fail by 10" threshold was reached. In general, AOE spells would inflict conditions one step less severe than single-targeted spells of the same level.

With that revision, AOE spells would only have three possible outcomes (unaffected, affected, super-affected), instead of four. Still more complex than binary, but not in a painful way.