I basically view that whole series as a conclusion in search of data. He set out to prove that the 4% rule doesn't work and found data to support his argument. (Plus, who's going to read 25 blog posts that state "4% WR still looks pretty good"?)
It's still a good thought exercise, and I think everyone should understand the underlying concepts. But some/most of them are trying to be WAY too precise and the exact methodology is too rigid to be particularly useful. I especially don't think that reading 10,000 words on whether your WR should be 3.47% or 3.56% is necessary, since as everyone in this Post-FIRE forum can tell you, not a single retiree withdraws or spends the exact same amount every year. A handful of basis points in either direction is going to have basically zero effect on whether your retirement is successful or not.
Jesus. We had people like that in engineering school. When comparing answers on assignments wanting to know what you got to the 6th decimal place. Forget that the huge assumptions on the way like whether you approximated a football as a sphere or a cylinder with throw even "correct" answers off by 50%.
A teacher then laid it out. In the real world, you run tests, you get data from sensors, you make best guess decisions and vast approximations, and you might have 1 significant figure confidently, and often only a certain order of magnitude. Data points will not line up in a nice curve and you can only make broad statements of correlation.
These models are not exact. Despite historically 3% never failing, the "true" failure rate is >0 because it's possible the future will be worse than the past. I'd go so far as to say that the true failure rate (even though it's unknown and hard to model) of 3% would be unacceptable if you were building a bridge.
Do you want financial certainty you can build a bridge on? Have 200x your annual spending divided amongst a few different asset classes. Maybe 50x in gold, 50x in T-Bills, 50x the market, 50x in RE. That way even if 2-3 entire financial classes shit the bed, and the last one offers 0% effective returns, you're still good for 50 years.
1.5-3% is what I'd call "almost certainly fine, but you never know" bracket. For this luxury you have to work 2-3x longer than someone in the 4-5.5% bracket with no better quality of life.
Similarly 3.5, 4.5, hell 5.5% are all in about the same ballpark. "You can expect to be fine, but keep on top of it and if you start to get much below your starting amount, maybe return to work"
Getting up higher to like 7-8%, "You're a gambler, give it a go, but probably won't work out for you but maybe?".
It's sort of like measuring the weight/displacement of a 100,000 ton ship to the gram, and calculating the margin of safety as to whether it floats or not, also to the gram (or Milli-Newton). The point is, if you're safety margin in anything depends on a factor 1/1,000,000,000th one way or the other, it doesn't matter which side of the fence you're on, you're already fucked as day-to-day uncertainty will rapidly eclipse that.
People are worried about a fraction of a percent on their WR, which is literally like $50 a month. Something as trivial as buying a new BBQ, replacing a stolen bike, or hell big things like getting hit by a car in Thailand make calculations to that degree of precision worthless.
For me the absolute best insurance policy is that I'm a thrifty, intelligent person who's navigated several 1-2 year breaks from work, and just started another which might be the last one. I'm probably fine, but I'll keep on top of it, and if not, I have lots of options.