Why crossover signals arrive late on daily charts

[FXS London Scalper]

Moving average crossovers are among the first indicators most traders learn. Their main weakness is lag: by the time the averages cross, a large part of the move may already be over, and in ranges they produce frequent false signals.

A worked example: a 20 and 50 period crossover on a daily chart may signal a trend change weeks after the turn. If the trend is long, the signal still captures most of it. If the trend is short, the signal may arrive near the end.

This means crossovers tend to work best in markets with long trends and fail in choppy conditions. A filter for trend strength, such as ADX or the slope of a longer average, can reduce whipsaws.

A method to evaluate is to test several period pairs on different pairs and timeframes, noting the share of profitable signals and the average win and loss sizes.

The indicator is not useless, but its role may be as a trend filter rather than an entry trigger.

Do you use crossovers, and how?

[Amanhasnoname]

I used to run a 20 and 50 crossover on daily charts and was frustrated by how late it felt, so I tested it on four pairs over about six years. The signals did catch most of the large trends, but the average entry came after roughly a third of the move was done, and in ranging periods I counted nearly two false signals for each real one.

What improved things was changing the role of the crossover. Instead of buying on the cross, I used it to define the direction and then waited for a pullback toward the shorter average on the four hour chart for the entry. That gave me a tighter stop and a better price, though I took fewer trades.

I also added a slope filter on the longer average. If it was nearly flat, I ignored the cross completely, and that single rule removed many of the whipsaw trades. The win rate only moved from about 38 to 43 percent, but the average loss shrank, which mattered more.