Your forgetting a very important factor in reviewing crit builds: Luck.
You can't quanitfy luck outside of running averages or running simulations, at which point you'd need to know how the game system actually works in the code or have a very high sample size from the game*. Since no game is ever going to have actually random results as one might expect them unless it enforces the percentages we see over x amount of rolls, it is entirely plausible that my estimations are more generous to Champions than the game is by making the player's effective crit rate be higher than stated. Or by making it lower than stated, because then consistent benefits become even more important.
*Sidenote: Your personal anecdote isn't a good sample size, unless you're logging crit rates as you play. Human memory is very prone to selection bias, so someone who is having fun with a Champion, or already thinks the subclass is good, is likely to remember more crits that mattered and forget more of the times they needed a crit, but didn't get one.
The reliance on luck means that Champions builds are unreliable at best, and the diminishing returns of their small increase to crits means that the same build with a different subclass is likly going to provide more benefits. Battlemasters and Eldritch Knights aren't just doing almost as much avg crit dmg on a crit build with a large enough sample size, but they also provide benefits that are consistent.
Champion is really bad in tabletop and the way BG3 implemented crit stacking makes them worse because it introduced diminishing returns on top of not being good to begin with. It's a wasted subclass that could've been replaced by almost anything else, or been given a proper homebrew treatment, like Beastmasters got.
You want a martial crit build? Build with Paladin, not Champion. You might crit slightly less, but at least Divine Smite is a mechanic that can make those crits worthwhile.
