I started with 2E, under a rather process-sim-style Dungeonmaster, so I never even learned about the whole poison-that-deals-damage-but-never-actually-hit scenario until fairly recently, on these forums. Going back to some of those early Gygax quotes, it still seems out of touch with my experience as to how the game is actually played.
The example that comes to mind for me is when the fighter is chained to a wall, and a dragon breathes fire on him. Gygax explained that it would be impossible to dodge the fire, if you were actually bound to that wall, so a successful saving throw vs breath weapon meant there was a weak link in the chain which the fighter subsequently broke before taking cover behind a nearby rock.
Which honestly baffles me, because I can't imagine anyone actually running it that way. While the dragon was heading over, I'm sure the fighter was already struggling to break free, and any chance to escape would be tested before the dragon gets there. The chance to break free would be based on Strength (either a Strength check against a set DC, or a bend-bars/lift-gates check), rather than the chance to save against breath weapon. If the fighter didn't already escape before the dragon breathes fire, he would get a saving throw anyway to see if it can dodge the brunt of the blast (possibly with a penalty for being chained up), or maybe the DM would say that no save is possible, but you wouldn't get the actions somehow conflated together into a single event. Positioning and state-of-being-chained-up-or-not are too important to the state of the scenario to be left to the whims of how you describe other things that are going on.
This is the point you are missing though - the fighter DID make his saving throw. He does take half damage. Now, you have to find some manner to explain that. This is a narrative that fits what the dice are saying - and it's interesting and makes for a fun game. After all, if we leave the fighter chained up, then, well, we might as well simply declare him dead since he cannot escape.
The whole point here is that the narrative has to fit what the dice say, not the other way around. The dice say that he only took half damage - why? Well, that's up to the DM.
This is why I don't really get why people try to use D&D as a simulation. It never really has been one. The fighter falls off a 50 foot cliff and walks away. How? Well, wouldn't it make more sense to adjust the fiction slightly so that that fall is now believable? Maybe he hit a few tree branches on the way down. Did the DM place trees there before? Probably not, but, then, DM descriptions of scenes are hardly so precise anyway. Having a few trees in the way that the DM simply hadn't described before since they weren't relevant is likely a lot more acceptable than having our fighter Wile E Coyote his way out of the crater he just made after falling off the cliff.
Abstractions are abstractions for a reason. They are stand ins for whatever is actually happening in the reality, but they are not the reality itself. A plus sign has no existence outside of a math question. We know exactly what it means, but, when I put an apple down on the table and then put another apple down, there's no magical plus sign that appears anywhere. It's an abstraction that we use to mean putting things together.
For an abstraction to be a simulation, it has to model events in the same way as the plus sign. When I see 1+1 on paper, I can visualise exactly what is going on in reality - one thing has been placed with another thing and now I have two things. When I roll a 15 to hit and deal 12 damage, there is nothing to visualise. There is no simulation here. It's no different than the old Final Fantasy games where your sprite jerked forward on the screen and a -X appeared above the enemy.
Putting it another way, using the D&D rules, prove to me that that's not what happens in a D&D world. If the rules are a simulation, that should be an easy thing to do. Show me how the rules preclude Final Fantasy style combat where negative numbers flash above your enemies after a successful attack.