You see the big downward trends in the stocks in 1998 and 2007?
Now pretend you invested in 1997 or 2006 instead of 1973. All of a sudden the graph would be transformed to show Gold as the clear winner. The guy's clearly an amateur, even as someone who has no interest in stocks whatsoever, never take advice from someone like this.
If you invest at the wrong time you lose money. The problem with graphs like this demonstrating the 'long-term' is that people conveniently pick a year that works for their point. The other common one is that people always pick just after the great depression to demonstrate stock market growth over time.
The real takeaway is to not try and time the market. don't just buy all at once. if you continually invest over time you will have much less volatile returns, and continually investing over almost any period the stock market has higher returns than the other asset classes mentioned.
This is never shown in these fancy charts, but any investor will see his bottom line affected by these costs.
Also, companies can be ejected out of S&P and replaced by more successful companies. These events are also routinely hidden from graphs.
If a company is ejected from the S&P 500, tracking funds tend to eject it as well. Also these charts tend to not be dividend adjusted, further increasing the benefit of stocks. Check out SPY; it actually has out-performed the S&P 500 (I believe due to dividends and what not).
These charts usually include dividends. When they don't they are very crappy. In many international stocks dividend is too substantial to be possibly missed.
Shows the S&P 500 clearly outperforming gold from 1973 to 2010. Of course, there were massive fluctuations; what this says to me is to be diversified (and diversification includes stocks).