by Helmut Schmidhofer
The variants of moving averages are so vast that you could almost say there is one for every trader that uses them.
Why are moving averages so popular? Because they automate trading decisions.
First, you have to decide which periods you want to average. You want to find the MA that hugs the lows in an uptrend and the highs in a downtrend.
Sometimes, this objective can be met by moving the MA forward. A good rule of thumb is to advance the MA by the square root of the period, i.e. for 2 to 4 periods, advance by 2, 5 to 9 by 3, 10 to 16 by 4, 17 to 25 by 5, etc.
Alternatively, you could have two MAs, one based on the lows, and the other on the highs. A downtrend has ended when the high MA is penetrated, an uptrend has ended when the low MA is penetrated. A valid penetration requires that the price has also moved beyond the previous period's range in the direction of the penetration.
If the low/high MAs are moved forward by the square root of the period then a penetration by a single price tick can be taken as a valid signal.
Some traders feel that recent data is more meaningful than older data, so they apply weights to the data. For example, a 5-period simple MA is the sum of the last five data items divided by 5. To calculate a 5-period linearly weighted MA, you multiply the first data item by 1, the second by 2, third by 3, forth by 4 and last by 5, add those numbers and divide the sum by 15 (1+2+3+4+5).
There are countless more moving average techniques. Each one suffers whipsaw losses and returns huge profits when the market trends. Trading with moving averages is the closest you will get to the rule "cut your losses and let your profits run".
Most traders are thrown by the fact that MA trading produces many small losing trades for every one big winning trade. It is easy to get disheartened after a number of losing trades and then find oneself out of the market when the compensating winning run happens.
It would be nice if every penetration had a label that announced the likely nature of the ensuing move. We could then avoid responding to false penetrations. Such labelling is possible with net-change oscillators, which are discussed next.
Back to simple moving averages Proceed to net-change oscillators
|
