Sorry, but your logic is wrong.
You're making the classic mistake of overvaluing direct personal experience, and because of that not correctly assessing risk. Very common, very human thing to do.
So, a beginners guide to risk analysis.
There are three main things to consider
1) The probability of a particular failure
2) The probability, given that failure, that we'll have a bad outcome.
3) The cost of mitigating that failure mode.
In our little example here, failure mode is the clip coming off and the chain separating.
The possible bad outcomes range from the bike coasting to a halt on the side of the road, necessitating a trailer ride home, to the chain cracking the crankcase and basically destroying the engine, to you getting rear-ended and dying, because it happened in heavy traffic with a car right on your tail.
The mitigation cost is the extra effort it takes to use a rivet link instead of a clip.
The crucial mistake you're making is assessing the failure probability based only on your own experience. Given that this is a relatively low probability event, that will always cause you to underestimate, because you have zero occurrences.
I have, between racing and spending many hours in a bike service shop, a lot bigger sample to draw on. I've never had a chain fail myself. I know many, many people who have never had one fail. But I also had seen a lot chains that did come apart, and I've seen some bad consequences as a result.
The other thing people get wrong is thinking that any failures of clip link are due to improper installation. That's something that I can't refute with absolute certainty, but consider this: I've had a clip come off (RTV held the side plate on, luckily). I know at least two professional mechanics that have had it happen. So, either it can come off when properly installed, or it's possible for even a pro to get wrong. Statistically it's basically the same.
Finally, we get to mitigation. It cost very little, in either time or money, to use a rivet link instead of a clip.
In the end, we have this.
A low, but not negligibly low, risk of failure. A possibility of serious negative consequences of that failure. And a cheap and easy way of virtually eliminating that risk.
Add all that up and to me, using a rivet link is a no-brainer