False signals on a trading chart usually reverse fairly quickly, putting you back in the trade in the right direction, but in the meanwhile, you take a small loss, called a whipsaw loss. Whipsaw refers to the whipping action of the price quickly moving through the moving average in both directions, resulting in a series of back-and-forth trades. Whipsaws occur in even the best-behaved trend and are common in a sideways market where traders are indecisive about trend direction.
Whipsaws have a pernicious effect on your profit and loss statement in two ways:
When trading a trend-following technique like the moving average crossover, you make most of your gains by riding big trends, and you accept that gains are going to be reduced by the occasional whipsaw at reversal points, sideways periods, and any spiky outlier. But if your big trends also contain whipsaws, you end up overtrading, which is to make a lot of trades for only a small net gain or loss.
Overtrading almost always results in net losses because on every trade you have to pay brokerage commissions and fees, which reduce profits and raise losses.
Instead of using the raw crossover of price and moving average to generate a buy/sell signal, you can set up additional tests, called filters. If the crossover passes the filter tests, chances are it’s a valid buy/sell signal and not a flash in the pan. Filters come in several varieties, and you can apply any or all of them to reduce the number of trades. Note that filters may delay entry and exit, and therefore may reduce total gains while reducing whipsaw losses.
Consider the following filters:
Time: The close has to remain above (or below) the moving average for an additional x number of periods after the crossover date.
Extent: The price has to surpass the moving average numerical value by x percent of the price or x percent of some other measure, such as the trading range of the past y days.
Volume: The crossover has to be accompanied by a significant rise in volume.
Extreme sentiment: In an uptrend crossover, the low has to surpass the moving average and not just the close; in a downtrend, the high has to be under the moving average, and not just the close.