Ask Colin

If I use a 21-week exponential moving average crossover for entries and exits (with time filters) I get a lot of false starts, but losses would have been greater without it. What do you think?

When I first learned about moving averages years and years ago, I spent a lot of time playing with them and tested lots of systems based on them on lots of charts. Some of my conclusions are:

Even with any filter I could think of, moving average crossovers were a far from good system. They work fine in a strong bull market when stocks trend strongly and clearly. However in sideways and down markets, they give too many false starts as you say.

It does not seem to matter what type of moving average I used SMA, EMA or WTA, the results were similar. Everything I did confirmed the well-known idea that trend following indicators only work as trading systems in trending markets. The rest of the time you get whipsawed to death.

In a few words - trading just is not that simple, although I understand a few very expensive black box computerised trading systems may be based on such a naive idea.

As you have observed, they will help you cut losses. However, there are lots of ways you can set stops and this is only one. It is nothing special in my experience, but definitely better than nothing.