Actually it's the opposite. 1e had attack charts. 2e progressed to THACO. 3e simplified to high AC and bonuses to classes that matched THACO but were easier. 1e had a few background skills that you might get one of. 2e had proficiencies, but once you had them you pretty much never got better or worse. 3e went to a better skill system, but with numbers that were too high. 5e simplified while keeping skills that could be improved. Class mechanics. Spell mechanics. Save mechanics. And more. All went through progressions that improved them.
All editions and versions of D&D prior to 3e have basically the exact same combat system, with a few slight tweaks or elaborations. 2e's THAC0 produces the EXACT same numbers as 1e's combat charts (there is an offset for certain very high ACs and the stepping for fighters is +1/1 instead of +2/2, but 1e offered that as an option). So, no, things pretty much stayed locked into exactly the same system. Even 3e's system has basically the same sort of model, attacks vs AC with a to-hit number followed by a damage roll, and save numbers for avoiding certain effects. All versions use a d20, etc. There was some change at the 3e break in terms of how the order of actions was determined, but that is really the biggest change in D&D combat. Fundamentally the model has remained unchanged for 40 years.
1e and 2e both had the optional background skills and NWP mechanics (not introduced into 1e until OA officially). This is a sort of incremental evolution, but note how these systems are always entirely optional until 3e, and all through AD&D they remain largely unworkable due to an adherence to a very limited type of design. Even 3e's version has problems based on its inheritance from NWPs. Only in 4e do we see some real evolution in the skill system, much of which has been subsequently undone in the name of 'tradition' (5e's system is somewhat of a mishmash of 4e and 2e influences).
I am OK with the idea that there were 'improvements', and if you compare 4e to 0e there are some significant differences, but even 4e is still working largely within the paradigms and set of mechanical tools established at the earliest times in D&D. Meanwhile other games have introduced entirely new, or radically restructured, mechanical solutions which D&D will obviously never be able to absorb due to its established structure (and the unwillingness of many players to change how they play).
Once you establish mechanical systems and subsystems it is very hard to go back and do something really different and potentially better. So creating poorly thought-out approaches to some aspect of the game and then expecting that to simply 'evolve' seems like a poor strategy to me. Most likely that element will remain almost the same, possibly being subject to revision at some edition break, but almost always with a lot of pushback and controversy.
Now, if you are talking about home brew, well obviously there's less of an issue with making ad-hoc changes where only a hand full of people use the system, but even then you may find it difficult to change.