Thanks for reminder but on hold. I got sidetracked into Inside Day strategy studies arising from another thread but I have not discounted your system for a bit more study.
I used to use averages until not so long ago. Now I've packed it in because I believe that they are unnecesary. What I do now is look at daily bars and come to a decision as to whether price is trending and which way. Then I look at 60M and 30M for possible turning points. Then I go to 10M and try to find the high, low for the morning, because I am not here in the afternoons.
If I am making a profit by 1330 I decide to take it or let it run at breakeven until 2030, when I get home. Then I decide whether to close or hold overnight. My main aim is to get into the daily trend and stay there for as long as possible.
No averages involved, no indicators, just a search for a turning point as soon after 0800 as possible. My stops, lately, have been very close. Right now, Monday, I am out with an 8 point loss.
All these decisions are very personal. No rules involving other peoples systems. Just my own experience as to where a line should be drawn. A horizontal line at a peak or trough on an index like Footsie is so approximate as to be useless, in my opinion.
If I am making a profit by 1330 I decide to take it or let it run at breakeven until 2030, when I get home. Then I decide whether to close or hold overnight. My main aim is to get into the daily trend and stay there for as long as possible.
No averages involved, no indicators, just a search for a turning point as soon after 0800 as possible. My stops, lately, have been very close. Right now, Monday, I am out with an 8 point loss.
All these decisions are very personal. No rules involving other peoples systems. Just my own experience as to where a line should be drawn. A horizontal line at a peak or trough on an index like Footsie is so approximate as to be useless, in my opinion.
This is johnson please send daily forex signals by mail: [email protected]
MA crossovers can be profitable
<ignores bizarre last post>
I'm with amnonco - I've proven to myself that MA crossovers can be profitable.
Urban Jaekle demonstrates step by step in his book "Trading Systems" (http://www.amazon.co.uk/Trading-Systems-development-portfolio-optimisation/dp/1905641796) how a simple, always-in-the-market moving average crossover system, trading an apparently arbitrary instrument (GBP/USD), delivers equity growth averaging about 5.5%/month (compounding) over a test period of about six years. That's with walk-forward optimisation (optimise on last 12 months, use resulting parameters for next three months, repeat quarterly), so there's no hindsight involved. Of course, the cleverness is in the periodic re-optimisation
I'm reproducing his findings as the first step in creating a collection of profitable automated strategies for myself, and I confirm it works. It doesn't seem hard to find combinations of parameters for an MA crossover strategy that are moderately profitable on any reasonably-liquid, trend-prone instrument.
Having said that - I don't dispute for a moment that there are higher-performing strategies out there!
<ignores bizarre last post>
I'm with amnonco - I've proven to myself that MA crossovers can be profitable.
Urban Jaekle demonstrates step by step in his book "Trading Systems" (http://www.amazon.co.uk/Trading-Systems-development-portfolio-optimisation/dp/1905641796) how a simple, always-in-the-market moving average crossover system, trading an apparently arbitrary instrument (GBP/USD), delivers equity growth averaging about 5.5%/month (compounding) over a test period of about six years. That's with walk-forward optimisation (optimise on last 12 months, use resulting parameters for next three months, repeat quarterly), so there's no hindsight involved. Of course, the cleverness is in the periodic re-optimisation
I'm reproducing his findings as the first step in creating a collection of profitable automated strategies for myself, and I confirm it works. It doesn't seem hard to find combinations of parameters for an MA crossover strategy that are moderately profitable on any reasonably-liquid, trend-prone instrument.
Having said that - I don't dispute for a moment that there are higher-performing strategies out there!
I think there is a tradeoff when we come to simplicity vs performance. You tend to find these simple systems are more statistically robust and belie the common wisdom that all trading systems only work for a short period of time. I haven't really come across (and maybe I'm the outsider here!) any high performance systems (e.g. MAR over 1.5 - 2.) which last more than a few years in walkforward testing.
The question for traders (aside from the majority who think trading mechanically could never be profitable, usually because of a desire for control and inflated sense of their analytical abilities) is:
a) do I make hay while the sun shines and trade this high yielding system for all its worth
b) do I use an old fashioned, boring old system like an MA crossover or Donchian breakout, put up with long periods without equity highs etc but be very confident that in 5 years, 10 years, 15 years time you'll still have a healthy average annual profit
I believe this is one of those decisions that can make or break a trader. My main caution with a) is that when that system stops working, you can give up a lot of your profits before you say thats it. A system stoploss will always have to be around 30 - 40% loss of equity (depending on position sizing) to stop getting whipsawed during normal drawdowns, which means when it does stop working, it really stops! On the other hand your boring old MA crossover will still be chugging along. Its a case of the tortoise and the hare!
I thought I'd just try something with option A. Lets look back at Futures Truth at the Top 10 systems since release date from 2004 (using the Internet Wayback machine) and see how they've performed since (the % is that given on their site at the time, which is annualised returns. I've put their total annual profit according to Futures Truth for each year since in brackets afterwards (i.e. 2005,2006,2007,2008,2009)
1. Mesa T-Notes 174.0% (-2350, -2575, 5408, -11679, 2446)
2. EuroTrader 139.3% (This system has vanished)
3. Anticipation 130.2% (This system has vanished)
4. R-Mesa 3 120.2% (This system has vanished replaced by R-Mesa 5)
5. Balance Point 110.1% (-8025, -24425, -5300, 63700, 550)
6. %C Daybreaker 104.2% (-10675, 21375, -8600, 37350, 27375)
7. I-Master 95.2% (-114425, -69775, 109525, 413935, -42726)
8. Market Rider 89.7% (vanished)
9. Dollar Trader for Currencies 85.4% (3511, -10064, 28713, 21540, -16366)
10. R-Breaker 77.6% (-20475, -7100, 42425, 8350, 9750)
There are definitely some profitable systems in there (I'm cautious with results from 2008 because that was such an unusual year). Many have vanished interestingly. You've got to ask yourself, at the end of 2006 would I still be trading I-Master, or R-Breaker?
I'd go with an option B approach. The one I'd have the most confidence in would be Dollar Trader - been around for 13 years at the time with profitable results, simple idea and fully disclosed. However I wouldn't have made the most money in the past five years. It will be interesting to look again in 2015 and see if that would have been the right longterm decision!
In the same way, I'd have a lot more confidence in five years that the dual MA would be making me money (on a diverse portfolio) than the latest genetic algorithm 15 indicator Ichimoku chaos based system
The question for traders (aside from the majority who think trading mechanically could never be profitable, usually because of a desire for control and inflated sense of their analytical abilities) is:
a) do I make hay while the sun shines and trade this high yielding system for all its worth
b) do I use an old fashioned, boring old system like an MA crossover or Donchian breakout, put up with long periods without equity highs etc but be very confident that in 5 years, 10 years, 15 years time you'll still have a healthy average annual profit
I believe this is one of those decisions that can make or break a trader. My main caution with a) is that when that system stops working, you can give up a lot of your profits before you say thats it. A system stoploss will always have to be around 30 - 40% loss of equity (depending on position sizing) to stop getting whipsawed during normal drawdowns, which means when it does stop working, it really stops! On the other hand your boring old MA crossover will still be chugging along. Its a case of the tortoise and the hare!
I thought I'd just try something with option A. Lets look back at Futures Truth at the Top 10 systems since release date from 2004 (using the Internet Wayback machine) and see how they've performed since (the % is that given on their site at the time, which is annualised returns. I've put their total annual profit according to Futures Truth for each year since in brackets afterwards (i.e. 2005,2006,2007,2008,2009)
1. Mesa T-Notes 174.0% (-2350, -2575, 5408, -11679, 2446)
2. EuroTrader 139.3% (This system has vanished)
3. Anticipation 130.2% (This system has vanished)
4. R-Mesa 3 120.2% (This system has vanished replaced by R-Mesa 5)
5. Balance Point 110.1% (-8025, -24425, -5300, 63700, 550)
6. %C Daybreaker 104.2% (-10675, 21375, -8600, 37350, 27375)
7. I-Master 95.2% (-114425, -69775, 109525, 413935, -42726)
8. Market Rider 89.7% (vanished)
9. Dollar Trader for Currencies 85.4% (3511, -10064, 28713, 21540, -16366)
10. R-Breaker 77.6% (-20475, -7100, 42425, 8350, 9750)
There are definitely some profitable systems in there (I'm cautious with results from 2008 because that was such an unusual year). Many have vanished interestingly. You've got to ask yourself, at the end of 2006 would I still be trading I-Master, or R-Breaker?
I'd go with an option B approach. The one I'd have the most confidence in would be Dollar Trader - been around for 13 years at the time with profitable results, simple idea and fully disclosed. However I wouldn't have made the most money in the past five years. It will be interesting to look again in 2015 and see if that would have been the right longterm decision!
In the same way, I'd have a lot more confidence in five years that the dual MA would be making me money (on a diverse portfolio) than the latest genetic algorithm 15 indicator Ichimoku chaos based system
lbranjord
Well-known member
- Messages
- 453
- Likes
- 15
I don't believe MA crossovers are profitable. The market efficiently takes these back at some point. Based on your selection of data you may find a play where they are profitable, even years. But that is not sustainable. Eventually you would be back at zero. That's my opinion. I've tested these in forex for a few years now manually and automated. Just can't get a consistent result.
It's an informative visual on the stock chart however, but most traders are long gone by the time they cross back anyway.
It's an informative visual on the stock chart however, but most traders are long gone by the time they cross back anyway.
Adamus
Experienced member
- Messages
- 1,898
- Likes
- 97
I haven't done any research into MA cross-overs but I think they could be profitable, depending on the time frame. If not daily then weekly or monthly definitely.
I think there is also something in re-optimisation, although I haven't done enough research into that. I don't know whether 12 months optimisation period, every 3 months re-optimisation is the silver bullet or whether those numbers (12 and 3) would vary. My first thought would be to optimise the optimisation period, but I'm pretty sure that wouldn't walk forward.
I think there is also something in re-optimisation, although I haven't done enough research into that. I don't know whether 12 months optimisation period, every 3 months re-optimisation is the silver bullet or whether those numbers (12 and 3) would vary. My first thought would be to optimise the optimisation period, but I'm pretty sure that wouldn't walk forward.
MA crossovers are an indication of a possible change of trend . They should not be traded just because they have crossed. Pullbacks are the way I go and, as for exits, I get out when I think it's time
., normally when I think that there is a choppy session coming, ie. the lunchtime period. Crossovers are terrible when they whipsaw. Like all this stuff, averages are tools of the trade and I use them all the time but the most effective time to use them is when they are trending, not when they are crossing over.With a crossover anything can happen but I find them more reliable with the larger timeframes than the smaller. I posted a couple of charts on "Swingin' the FT 2010", a couple of weeks ago, but with the 4 hour frames one has to be careful with the timing because of the large value in the length of the bar.
This is all opinionated stuff but guys do go into a lot of detailed work and I can't, for the life of me, see the point. Averages are so simple to use, but stay away from small timeframes.
MA crossovers are an indication of a possible change of trend . They should not be traded just because they have crossed. Pullbacks are the way I go and, as for exits, I get out when I think it's time
With a crossover anything can happen but I find them more reliable with the larger timeframes than the smaller. I posted a couple of charts on "Swingin' the FT 2010", a couple of weeks ago, but with the 4 hour frames one has to be careful with the timing because of the large value in the length of the bar.
This is all opinionated stuff but guys do go into a lot of detailed work and I can't, for the life of me, see the point. Averages are so simple to use, but stay away from small timeframes.
A. you can trade MA x overs.
B. "longer/shorter" time frames are subjective to a specific trader. for me longer means 300 days and shorter means 20 days. i don't talk in hours.
as adamus mentioned, you trade MAs when the noise is minimal and thus the crossover is meaningful. if you're looking at hourly charts, of course you can't trade MAs.
from my experience and testing, really long MAs are profitable and there's no real need of optimization (since you cannot implement that forward).
just pick an asset, try to use 20-60 day MA as short and 100-300 day MA as long, go long only, or long&short, and check the results. you'd be surprised.
There you go. There's more than one way to kill a cat.
lbranjord
Well-known member
- Messages
- 453
- Likes
- 15
In FX, trading daily charts based on any ma pair seems to lose or maybe break even. If they do manage to gain pips in the end, it is less profit than a bank money market account.
I like them on stocks though!
