Mark Hulbert at Marketwatch writes about the 50 day moving average (link). In the bigger picture, there is this:
As I’ve written on prior occasions, something appears to have happened to moving-average trend-following systems in the early 1990s that left them far less able to beat the market.
In fact, according to Blake LeBaron, a Brandeis University finance professor who has extensively studied a number of different technical trading rules, moving averages of various lengths stopped working in the early 1990s, not only in the stock market but also in the foreign-exchange markets as well.
I can guess at "what happened." The Internet became more popular as did technical analysis packages, and then free websites offering charts. Pre-Internet, moving averages were the realm of a few that took the time and expense to have a data provider and a charting service. Once charts became popular, the most popular indicators became less valuable.
With all that I still look at the 50 and 200 day moving averages. Many others look at the 10, the 20, the 150 day. While the indicator is no longer gives a high probability chance of success, if it is widely followed, it can serve as resistance, support, and a whipsaw level. In human terms, I look because so many others are looking.
It is a cautionary tale about indicators that do work. If everyone starts using them, they tend to stop working. I cite this a lot when writing about seasonal indicators, because the calendar is easy to track, and easy to try and jump.