For me, most of them. TSR D&D, WoD (VtM1e, CtD1e (the cardboard system is NOT inscrutable, really fairly simple), and so forth. Revised stuff actually pushed Storyteller that beanbag of math-phobic kludge of mechanics to breaking), Blue Planet, CoC (vs. 7e; earlier 6 editions are close to the identical), Fading Suns (simply benefit from the experience, it really works higher in observe than you assume — a typical revelation like a variety of early D&D-isms), L5R (although I’ll give allowance for 3e-based LBS, as that’s its “1e”), SW WEG, Shadowrun (sure, even 1e looseness & 2e punchbowls of cube is healthier than its devolutions — a lot devolution!), Mechwarrior & WH1e (however solely in passing familiarity; I performed in stated video games, by no means ran them), Deadlands1e (sure, even with the poker sport mechanics), Traveller1e and even MongTrav1e (sure, even demise at chargen is extra a comedic snicker than a play nuisance of
ALL THE MATHS, or setting flipped over), early Palladium was nonetheless inside manageable sanity ranges…
Hard to think about any system that received higher over time, now that I give it some thought. 
I’d say GURPS hit its apex someplace round 3e, although I can see arguments for 2e as properly. 4e is… a helpful compendium for individuals who already know what they need out of GURPS. That one I might see received higher over time; I couldn’t argue it in addition to others, however the sourcebooks did get far more helpful together with higher manufacturing over time.
Another one I’d concede can be CP, as a result of I like CP2020, however I haven’t got sufficient expertise of the sooner stuff to stake a tough stance on that. But then one might depend CP2020 as its personal 1e, and I’m not terribly gained over by the most recent CPProjectRed as of but.
