Since my last post on this (See here)
The main thing was the usefulness of having a trend filter. David Vardi suggested that trend following systems work best with either a filter or a percent exposure money management based on the longer term trend.
For example, if trading a 100 period channel breakout, the theory is that you will get better returns if you use the 200 period donchian as a filter. If the last breakout of the 200 period was to the upside then be stronger on your long trades if the last 200 donchian was to the downside, be strong on your short trades.
Does this sort of thing work?
Let's find out.
The rules were as follows:
- Long if there is a close above the 55 period donchian (D).
- Close if close below 20 period D
- Stop based on 2 times the average true range, trailing.
- Shorts, the opposite of longs.
- Stop 2% max of equity.
The filter for longs - go long only if the last break of the 100 period D was to the upside. If not do not go long (vice versa for shorts). In other words, compare the last 100 day high and the last 100 day low. If the 100 day was more recent then go long.
Forex markets tested on daily charts:
GBP/ USD, AUD/ USD, USD/ CAD, USD/ JPY, EUR/ USD, EUR/ JPY & GBP/ JPY
Results using filter:
Profit factor: 1.5*
Max DD: -12.79%
S/R: 46%
Results without filter:
Profit factor: 1.17*
Max DD: -16.5%
S/R: 39.47%
Chapeaux to mr David Vardi then because it seems the filter does make a real difference. Profits, Draw Downs and Strike rates are all improved with the filter.
But what about lower time frames?
This time I used the same rules and same filter, except I used the 20 D to enter trades and the 10 D to exit trades.
Results using filter:
Profit factor: 1.46
Max DD: -14.5%
S/R: 40.65%
Results without filter:
Profit factor: 1.12*
Max DD: -19.74%
S/R: 38.17%
Using the Price Rank function as a filter:
I wondered if I could use my new favourite tool, the Price Rank tool as an additional filter.
So here goes, as above, except for longs the Price rank (252 periods) has to be above 0.5. For shorts, it has to be below 0.5
55/20 Results using additional price rank filter:
Profit factor: 1.56
Max DD: -12.78%
S/R: 46.44%
20/10 Results using additional price rank filter:
Profit factor: 1.42
Max DD: -14.62%
S/R: 40.63%
Not much difference on the 20/10, but it certainly makes a difference on the 55/20. Food for thought.
Profit Factor:
*(PW * AW) / (PL * AL)
Where
- PW = Probability of a trade being a winning trade
- AW = Average win size
- PL = Probability of a trade being a losing trade
- AL = Average loss size
Basically anything above 1 is profitable. Anything below 1 is loss making.
