A well-designed matchmaking system should leave you with a 50% winrate. That's how Elo and Elo-like ranking systems are designed to work.

If your Elo is climbing over time, that's (probably) good progression. If your Elo stays stagnant with a 50% winrate, that's you not getting good. With the caveat that how "good" a particular Elo is can change over time, since the starting Elo is by definition the average. If existing players are learning to be better, or worse players are quitting the game, or better new players are joining, remaining at the same Elo might mean you're actually improving; you're just improving at the same rate as the average player. And conversely if more casuals join or hardcore players quit, the average should go down.

Then there is a balance patch and all the builds you once knew are different and you have to play different

I definitely found that to be the case in AoE4, and to a lesser extent in AoM, but never in AoE2. I think partly it has to do with how mature the game and its balance are. When they get to a stage where fewer big changes are being introduced at once, older builds are probably still viable.

I think it can also be easy to get caught up watching the pros' "Patch X Civ Y build order" videos, thinking you have to watch and follow the latest patch's BO, when really this BO is often just an extra option on top of BOs that worked in the previous patch and still can work. But the pros gotta put out that content to appease the algorithm.

I haven't played much AoE2 in quite a while, but I do love playing Lith. Prefer leitis to knights, if I can get them.