The question they asked
Cooley, Hubbard and Walz (1998), often called the Trinity study, asked how often a fixed initial withdrawal lasted 30 years on overlapping historical windows, for a few stock/bond mixes. A 4% initial rate on a stock-heavy mix lasted in most of those windows. That is a historical success table. It is not a probability that your retirement lasts, and it is not a 50-year result.
What we did not reprint
The Trinity study calculator asks the same kind of question on 1,000 seeded paths you can edit. It is not their tables. If you want the paper Killion actually ran — 5,000 regime-switching lifetimes, flexible versus rigid 4% — that is is the 4% rule still safe. If you want the rate to move with age, that is withdrawal rates by age.
How to use a success table
Compare two rates on the same engine. Do not take a 95% from 1998, a 92% from a blog, and a 88% from this site and average them. Different samples, different inflation treatments, different failure definitions. The reading-Monte-Carlo guide is the hygiene.