...If something is 95% effective, there's a reasonable chance you'll be unlucky and have it not work for you.
And placebo effect is about 30%. That is, 30% of people will improve with any treatment, even one that can't possibly be any use (classically, sugar pills). So if you have something, and you get improvement, it's quite likely the placebo effect.
(18 moths ago I had a real problem with my elbow on rides over 100km. It would lock up straight so I couldn't bend it without spending several minutes working on it with my other hand. I seriously considered getting a bike fit, but decided not to as it was too close to a very long ride so didn't want to risk a new position without time to get used to it and also because there was not much in google about elbow pain with cycling so it seemed likely my problems were outside the scope of most bike fits. Anyway, 2 days into my long ride I found the problem had suddenly gone. If I had gone to a bike fit, I'd be claiming it worked brilliantly, yet it would have been just a coincidence)
Firstly, 89.47527% of stats are made on the spot.
Secondly, jokes aside and using your figures only, you don't seem to have grasped how stats work? 5% chance is hardly termed reasonable, when 95% is saying it will work. There's a chance, no question, but you suggest it's likely when the fact is that it's 19 times more likely to work than not. If it's a serious operation we were talking about, you would take those odds!
Quite likely at 30% is a fair term, but you're two and a third as likely not to experience the placebo effect.
The elbow thing is interesting, as any bike fitter would agree that you should not cycle with locked elbows for 100km. There's no surprise that you experienced this. The next ride, you probably cycled for longer with more bend at the elbow. A proper bike fit would resolve this and take the chance out of it.
As far as bike fits go in general, it depends on how switched on you are in setting up your bike up to start with. If you just ride it like set up in the shop without them seeing you or as a rough guide stood next to them, then you are considerably less likely to have the optimum fit than if the fitted you properly.
If you look online at the things that you need to adjust to get close to a good fit, and work at this over time, then you are considerably less likely to see such big gains as the guy in the first example. Stands to reason really?
What I find amusing, is people saying bike fits are pointless, when they have actually done many of the adjustments themselves already! That's irony for you at its best!
If you have adjusted your bike and still riding with pain, and nothing has changed to you physically or the bike, then it's not set up right, or the wrong size, or you're simply not doing enough riding to get used to the desired position.