This automatically assumes you literally hate what you do. I could say that I see owning real estate as an automatic failure because I would literally hate dealing with it. It doesn't make it true, but it makes it true in my case. My issue with the 30 year studies vs. the 40 & 50 year simulations is as you start to get into larger and larger years for simulation purposes the number of unique combinations goes down. So we have no where near as much historical data on a 50 year withdrawal rate as compared to a 30 year withdrawal rate. It also assumes the US will continue to be a GDP super power or that you will know when to invest in any new emerging markets that become the new GDP grower that the US has been over the last century. Could a 4% withdrawal rate work over 50-60 years, sure, but I wouldn't rely on historical simulations of that considering most of the data used in the simulation is not even double the time period you are trying to simulate. We need to remember that the life expectancy of healthy adults continues to increase. I am in my early 30s, there's a strong chance that the average life expectancy for people my age as long as they are healthy at 60 is into the mid 90s. So if you retire planning on a 4% WD exactly and little no wiggle room at 40 you may need to plan on supporting yourself for 55 years and that is just assuming it is you that needs to be supported by that money. A married couple may have an average life expectancy between the two of over 100, which would mean you need to make sure that money lasts over 60 years.
Now you could say don't worry about it because you can always go back to work, but I'd rather do what I do and enjoy for a few more years than chance needing to go back to work in my 60s or 70s. To each their own, but I definitely wouldn't say working a few years longer is automatically a failure especially if you don't dislike what you do in the first place. You need to keep in mind that not everyone in this forum is aiming for FIRE as fast as possible, some are aiming for FI and a relatively early RE.
There's a couple of things that jumped out at me with your post I'd like to respond to.
1) From a purely medical standpoint this notion about increasing lifespans is wildly misunderstood. Ignoring for a moment the recent dip in life expectancies (mostly due to an increase in drug overdoses and suicides among otherwise young, healthy individuals), the increase in life expectancies is because fewer people are dying earlier
. When fewer adults die in their 50s or 60s the median life expectancy increases. Put another way, while the life expectancy of a healthy 60 year old has increased by roughly 3 years in the last two decades, the probability that they will reach 90 has barely budged. As a species our average (median) life span has increased, but our maximum lifespan is essentially unchanged.
There is no evidence that the life expectancy (by definition a 50% probability of surviving) of someone in their 60s will be their 'mid 90s'. Sorry, but the likelihood that either person in a marriage will live to see 100 is very small. Not today and not projected for our generation (you and I are about the same age). Currently a healthy 60yo white male has a life expectancy of 81.5. I'd love to think my wife and I will live to see 95, but the odds are heavily against it.
this doesn't mean that financially you shouldn't plan for your money to last as long as your maximum life span - you should. On that note...
2) While it is true that the available sample size (or 'number of unique combinations' as you call them) does decrease as you increase time periods, this doesn't mean that its not possible to test longer time periods. The most common way to do this is with Monte Carlo simulations where you re-sample (sampling with replacement) different time periods, which allows you to have very large and equal sample sizes even when the time periods are very long. Most people seem to use yearly time periods for MC simulations, but there's logic to sampling at 7 or 10 year time periods as well (cyclical markets and all...) An even simpler method is instead of viewing how many simulated portfolios 'failed' is to view what percentage ended with a higher real balance than they started with. There's a far greater probability that you will end your first 30 years of ER with substantially more then there is that you will have less money (in real terms) than when you first retired.
All of this assumes that the next few decades will be no worse
than the worst time periods of the last 100+ years, which brings us to...
3) The idea that the lost of the US as the largest GDP somehow equates with the eminent failure of the 4% 'rule'. There's several different facets illustrating why this need not be the case. For example, the ratio of US GDP to global GDP has been falling since its peak in 1952. There's the fact that only ~55% of profits from SP500 companies come domestically (a number that's also been steadily declining), and the idea that 'emerging markets' are better long term investments (historically they have not). Lower growth won't even limit the efficacy of a 4% WR per se
. For that to happen we would need more periods that were as bad or worse than what we've seen over the last several decades. Despite the 'great recession' and lots of articles published over the last 4 decades claiming 'the good times are over' there's little evidence showing that this is actually happening.