Version: 0.5.22b

From the actioneffect function in newsexsystem.gd

if values.tags.has('pervert') && (acceptance == 'good' || person.traits.has('Pervert')):  <-- So, if it's a pervert action and you ACCEPT it or are ALREADY PERVERTED
self.lust += lustinput
self.sens += sensinput
if lust >= 750 && randf() < 0.2:
actionshad.addtraits.append("Pervert")  <-- You can GAIN PERVERTED . . . which you might already have.
else:
person.stress += rand_range(2,4)  <-- Or get stressed out . . . 
elif values.tags.has('pervert'):  <-- However, if you are NOT PERVERTED and DON'T ACCEPT the action
self.lust += lustinput/1.75     <-- You gain more lust?
self.sens += sensinput/1.75  

Surely that's backwards.  If you;'re not perverted, you might gain perverted or just get stressed, but if you are already perverted, you just gain more lust and don't get stressed.

I too think that the perverted actions are the wrong way around. I did have a Unix-style patch here but I took to down, because:

1. The Tab character is a bit important in GDScript fragments (offside rule and all that), and the itch.io message board software (or maybe Chrome itself) doesn't really like it; all my tabs were expanded into spaces, which messed up the intended indentation. I could (and did) mangle my code fragment to overcome this, but it still wasn't pretty.
2. More importantly, my patch was buggy and made it all worse -- it stopped perv actions from building stress and eventually getting the Pervert trait.

If I ever find a good way to post GDScript fragments here, I'll do it again.

x < x/1.75, so you do not gain more lust. You don't gain stress, but i'll be fixing this.