I wonder what the longest duration has been, in the history of let's just say S&P500 or similar, between [a low so bad that people would say "sucks that I'm retiring now"] and [the point in time prior to that, where it had the same exact value but as an all-time high]. A decade maybe? So basically one would conclude that they could've begun taking withdrawals a decade ago with the same size of retirement account (well, except for all the contributions throughout that decade, and except for the extra decade of remaining life expectancy, and prevailing wisdom to reduce risk as retirement approaches, each extremely significant, but the point remains to some extent).

There was a study (https://en.wikipedia.org/wiki/Trinity_study) showing that you could count on a 4% withdrawal rate for retirement income over any 30-year retirement period, including periods spanning the Great Depression and other market lows. Most of the time you can count on more; that was the lowest across the entire study. And with compounding, the difference between "30-year retirement period" and "indefinite retirement period" is not much; withdraw a bit less, or be able to adapt to prevailing conditions, and your retirement portfolio will last indefinitely.

Are you talking inflation adjusted or nominal? Total returns?

I guess nominal, because not having kept up with inflation would be the main source of it sucking badly. I think of it like this:

You lose to inflation badly with uninvested cash. You get close to keeping up with inflation with near zero risk vehicles (CD or high yield savings account). You hope to at least keep up with inflation with medium-low risk vehicles (S&P500 or similar funds). You roll the decide after that (leverage, timing, stock picking, etc.).

> You lose to inflation badly with uninvested cash.

Details depend on what currency you pick (and what decade you are in).