I also feel it makes the masturbation scenes a game mechanic as well as eye candy. I think paperwork reset arousal each day, if that is the case I would change that to lowering it somewhat when she sleeps, but leave it at maybe 75% or something like that.
I think there was a reason for PW to do this. Often you do not run into problems with first ideas of mechanics, they will show up when you code it! Too many states, variables can be tricky, well not in coding, but to deliver the visual stuff for so many.
Let's say the player don't let her masturbate for what, five days? And, yes exaggerated, she's therefore ready for a 100 men bukkake gangbang with unwashed, stinking hobos. And the player played her more shy, what then? Is there no scene. Maybe I didn't make the point clear, but carrying over variables, over more and more days can be a pain the ass for logic, renders, dependencies etc.
So the advantage with the daily reset of the variable lust, he could handle the value within one day of the update, also with the renders you need for the different variants of a scene depending on her lust value (and that is just one variable within a day). Now think about two, three or more variables. You have to make some cut, it sucks I know, but otherwise the complexity will kill the project.
I am not saying you can't use the lust variable as the main "ressource" and drunk, stoned as addtionally outcomes of a scene, but keep in mind this needs to get tracked and somehow be logical. ANother example, after three days not touching herself, there is a possible scene in the subway. So you need (should?) take care for a groping scene not only for a parameter how drunk she is (maybe drving home after the cosplay modelling), maybe she also smoked some pot after the show with someone, but you to check her lust value, not only for the day or two but four, five. How many states of lust can she have, maybe three? Or five? So the scene could play: lust 0, drunk 0, high 0. lust 1, drunk 0, high 0. lust 2, drunk 0, high 0. lust 0, drunk 1, high 0. lust 0, drunk 2, high 0. lust 0, drunk 0, high 1. lust 1, drunk 1, high 0. and so on and on. And maybe the variance of two outfits? Doubles the possibilites instantly. Are there dependencies of previous encounters with the guys in the subway groping her tonight? Will they be rough or gentle, how far will they go, will they humiliate her in front of others because she is drunk 2, high 1, and treated them condescending in a previous event, does lust even matter in this case?
Better you use less variables that really matter the outcome of a scene after a choice (or maybe in future) then having many variables but you see no difference in the scenes!
Same for outfit changes. New outfits are always nice, and the dream would be depending on her "overall" lewdness. Yeah but some players play her not like a bitch, different scenes will be played different. So now you are on a point to have a certain scene, and the players have her on different lewdness. You plan to render those upcoming scenes with different outfits? -> project dead. The amount is too much!
So imho, I would say one overall parameter for her, like corruption (if it should be a corruption based game). So this can be used to "force" the players with every update within a certain corruption level (you know the minimum corruption players could achieve throughout the updates, and a maximum). And use things like drunk, high, arousal only on a daily basis for the events within a day. You could also isolate some event chains like the bathroom scenes. Still you would also need to track her overall state. It's kind of unlogical if she does a hobo gangbang in a dark alley but reacts prude if some of the brother peep on her. Oh so many things to keep in mind.
But what scenes we like and don't like is obviously just taste. I think he did quite well with the tools he had.
