With respect, your personal experiences are not data, and should not be treated as such to counter arguments.
A major problem we have is that nobody does regular surveys of RPG players to maintain reliable data. But, I'll turn back to the 1999 WotC market research for some idea.
In that research, the questioned folks about how many sessions of play they went through before "reset" - starting over with new characters. The results:
If you've been playing for
less than a year - average sessions before reset 8.8
between 1 and 5 years - average sessions before reset 12.9
over 5 years - average sessions before reset 19.6
So, for what we can call "new" gamers, the average number of sessions before reset was not high - 8.8.
While we don't know the distribution, the basic assumption would be that about half of all campaigns for those new players then lasted for fewer than 8.8 sessions.
For folks in the 1-5 year range, if they were playing weekly, the game lasted - on average, about three months. Maybe six months if they played once every other week.
Thus, games falling apart quickly, for whatever reason, was common. While this was a quarter-century ago (!) I think you'd need to make a pretty compelling argument to push the idea that this still isn't the case.
I don't think any individual's experiences - one way or another - tell us much of anything. Do some games fall apart quickly? Of course. Do we know how many? No.