The OP contained 2 very valid questions and this post has strayed far from them. The questions were:
... the whole system that we're basing our statement on are not that old. Sure, the stock market always goes up, but only in the last, what, 150 years? It is a baby as far as data is concerned.
...
What about major game changers ... ?
1) You are describing the "fat tails" problem in statistics, popularized by Nassim Taleb in "The Black Swan". In a nutshell, the problem is that we do not know the actual frequency of rare events based on a sample. This is because events with, say, a 1% chance of occurring are not likely to be picked up in a sample of, say, 15 observations, such as the number of non-overlapping 10 year histories of stock market performance, or the 5 observations of non-overlapping 30 year market performances. There could be many such events with 1% frequencies. Also, in the real world, outside of careful laboratory controls, millions of things happen every day for which there is no historical precedent. According to Taleb, traditional bell curve thinking probably doesn't even apply to big complex outcomes such as national collapses, terrorist attacks, or stock market returns.
2) Major game changers are fairly common if one looks beyond the history of US markets. Consider that in 1900, Argentina was a world-leading economy and the US and Mexico had similar GNP per capita and living standards. Imagine being a Turk invested in the Turkish stock market when Recep Erdogan overthrew the establishment there. Think about being invested in Venezuela, Zimbabwe, Hungary, Greece, etc. In my hindsight, the key factors that differentiated the US from the basket cases were (a) a stable and slow-moving political system, and (b) at least somewhat competent leadership. If those two assumptions ever fall into question, I see no special luck that would prevent the US from having the experiences we're used to only reading about, like hyperinflation, equities collapses, or 50% unemployment. The exceptions to my assumptions are post-1990 Japan and possibly post-2016 Britain, both places with stable governance and moderately competent leadership which have made disasterous policy decisions. The other examples might have been predictable for someone with the right perspective, i.e. Venezuelans who got out with their assets when Hugo Chavez came to power or anyone who shorted bitcoin last year.
The question is to what extent to diversify? My portfolio is about 90% stocks (index funds) based on the rationale that these assets have the highest risk-adjusted expected return and that shifting into bonds, commodities, real estate, P2P lending, notes, crypto, cash, art, etc. would probably - but not certainly - involve a reduction in overall portfolio returns.
Because we cannot empirically estimate the probability of fat tail events that would make any given hedge a good investment, Taleb would say we are flying blind with asset allocation. His own AA as a fund manager was something like 90% treasuries / 10% speculations like call options on the S&P 500 - an upside-only portfolio that reflects an expectation that the market is underpricing risk.
I did something similar, establishing protective puts on about half my portfolio last summer. It was a good decision, based on erosion of my 2 assumptions about the US market.