It's wrong to use direct statistical analysis to describe how stock prices behave for any really long period of time. Mostly because it is only valid to apply those methods when certain conditions exist, when the market might be described as being orderly, or if you really like mathematical puns, when the market is exhibiting "normal" behavior.
For the current stock market, those conditions have existed since 4 August 2011. Let's have a look.
Looking at the overall statistical trajectory of stock prices at this scale, and applying what we know about the fractal nature of stock price changes, the recent three standard deviation drop in stock prices since peaking on 19 September 2014 should be considered something of an early warning signal, as it suggests that the present state of order in the stock market is at a greatly increased risk of breaking down.
But there is no guarantee that it actually will, which is the curse of statistics-based analysis. You can simply observe the chart and see when similar early warning signals were sent without order breaking down, which would be indicated if stock prices either rose above or fell below the outermost (red dashed) statistical trend lines shown in the chart above, which are positioned at a distance of three standard deviations away from the central trend trajectory.
If you did want to make an investment decision related to the change in stock prices, you would wait until they definitively crossed one of those two curves to avoid the risk of acting too soon and losing out on potential gains while racking up unnecessary transaction costs.
For now, the more-than-two standard deviation shift is simply an indication that you ought to pay closer attention to what stock prices and doing and really ought to have some sort of strategy worked out for what you will do should the current state of order definitively break down (say if the S&P 500 suddenly dropped to close below 1850).
Because, when the conditions exist that make this kind of analysis appropriate, it can be an exceptionally valuable tool for making investment decisions when it matters most.
When you can still do something about it.
Update: 15 October 2014 9:41 AM EDT: You can't say you weren't warned. As Scooby Doo might say, "Ruh roh!"