What Jay R said, plus:

If you look at 1e, advancement really stopped at around 9th level. After that, you only got nominal increases in hp and attack scores. Sure, spellcasters *might* get spells if they had sufficiently high scores (and with random stat generation, most didn't). But really, advancement capped out around 9th level.

Combine that with the fact that you started running out of stuff to fight, and it started making sense to retire those characters at that point.

Keep in mind, too, that in the original campaigns, you'd have multiple characters. So retiring one just meant you played others, not that you had to stop playing the "adventuring" game.

Once you got rid of that soft level cap, and made the assumption that people would keep playing the same characters, and that you had to support the higher levels, a lot of the assumptions that worked in the default 1st-9th level range stopped really working.

The biggest problem with D&D is that it's built around a playstyle that not many people use today.

If 4e did anything right, it's to try and build a game specifically around how "most" people play, and make a game that gets rid of inaccurate assumptions. I'm not saying they *did* that effectively, to be clear. But I think the attempt was the right idea.