Technology

History is the wrong way to look at it. If I go to a roulette table and the last 10 balls have landed on red, that doesn't change the odds that the next ball will end on red. Assuming the table and ball are fair, no amount of history will change the probability of the next ball landing on red, so there will always be a negative expected return for any bets placed on a roulette table.

Investment grade securities are different. Businesses are expected to return a profit to shareholders, otherwise they will eventually go bankrupt. So there is a built-in expectation of positive return, regardless of the pricing history of the security. Buy-and-hold investors should always expect a positive return on a diversified portfolio, because, on average, businesses are expected to return a profit. Valuations can certainly fluctuate in the short and medium term separate from profits (valuations include future expected returns), but since it's not a zero-sum game, long-term returns are expected to be positive.

So no, 15-years is not enough, nor is any other arbitrary amount of time. Any expectation of future returns should be largely founded on the underlying fundamentals of whatever it is you're buying, and then modified by past returns to adjust the probability of returns going forward.

Stock valuations are a poor proxy for actual value, they're more a measure of market sentiment, which is why very short-term stock movements (esp. less than a year) are largely just gambling, because as A. Gary Shilling said:

Markets can remain irrational a lot longer than you and I can remain solvent.

So if you're trading good securities (i.e. companies with positive expected return) on a short-term basis, you're likely gambling, because you're betting on changes in market sentiment. But if you're trading good securities on a long-term basis, the underlying fundamentals should outpace short-term deviations from expected returns. My general number here is 10-years, but many financial experts go as low as 5-years in terms of investment horizon. The historical returns matter a lot less the shorter or longer your time horizon, and IMO are largely only important on medium terms (say, 5-15 years) because at that point we're looking at more systematic over or under valuations (and tools like the Shiller CAPE do a decent job of indicating that).