Why exactly do EMA crossover systems not work? | Page 20 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Help. EMA Crossover with different EMA cross for exit 0 replies

EMA crossover (ema 14 & 34) 4 replies

Please help code an EMA Crossover 12ema x 24 ema cross 1 reply

Systems, Systems and Systems... 1 reply

EMA Crossover Systems 1 reply

Attachments: Why exactly do EMA crossover systems not work?
Exit Attachments

Why exactly do EMA crossover systems not work?

  • #381
  • Jul 27, 2011 8:45am Jul 27, 2011 8:45am
I use 2 emas as only indicators on my charts, but i don't even use them as signals more like an "anchor" you know somehing to referce from to easier visualize. My signals are only based on price and SR either stragith lines,trend lines, boxes and triangles
The universe seems neither benign nor hostile, merely indifferent.
  • #382
  • Jul 27, 2011 12:54pm Jul 27, 2011 12:54pm
Quoting Custos
okay, you probably have no knowledge in portfolio management.
Custos, I started studying the currency market back in the 70's, when the Bretton Woods system was still in place, so I know a few things about portfolio management.

You are trying to obtain a smoother equity curve and reduce drawdowns by combining two or more trading systems, assuming your (hopefully profitable) systems are not too much correlated. This idea is as old as the sun, nothing new here, all I am saying is that a lot of traders think that if System A makes 25% a year and system B makes 25% a year then all they have to do is trade System A and B simultaneously and they will automatically make an easy 50% a year (2 x 25 = 50).
  • #383
  • Jul 27, 2011 1:40pm Jul 27, 2011 1:40pm
Quoting jzw
So I accept your point that it works, I just think in the real world no one would use it.
Here are the results of another simple trend-following system. It's called the Aberration system and it is one of the top 10 trading systems of all time, according to Futures Truth, and independent service that tests and ranks publicly offered futures trading systems. The second number on the right is the average profit per trade.

System tested between 1980 and August 2010 (30 years)

Japanese Yen $57,125 --- $921
Swiss Franc $72,000 --- $1,263
Canadian $19,489 --- $341
British Pound $19,018 --- $327
Dollar Index $62,420 --- $1,273
Mexican Peso $325 --- $10
Australian ($) $2,419 --- $43
Euro Currency $63,575 --- $1,271 (Euro did not trade until 2000 so I guess they must have used the ECU or the German mark as a proxy between 1980 and 2000)

Total number of trades : 419 (about 1 trade a month)
Totals $296,373, or $707 average profit per trade per contract.
($50 allowed for commissions and slippage)
  • #384
  • Jul 27, 2011 3:22pm Jul 27, 2011 3:22pm
Quoting Mr Breakout
Custos, I started studying the currency market back in the 70's, when the Bretton Woods system was still in place, so I know a few things about portfolio management.

You are trying to obtain a smoother equity curve and reduce drawdowns by combining two or more trading systems, assuming your (hopefully profitable) systems are not too much correlated. This idea is as old as the sun, nothing new here, all I am saying is that a lot of traders think that if System A makes 25% a year and system B makes 25% a year then all they have to do is trade...
well, your last post was directed at me, so I assumed that you meant that I don't understand that fact (probably you still do).

Anyways, I am not guessing here, I actually tried very hard to find systems that are complementing each other and also pairs that complement each other, so that I can achieve higher performances than if I would just use System A or B.

cheers
  • #385
  • Jul 27, 2011 3:46pm Jul 27, 2011 3:46pm
Quoting Custos
well, your last post was directed at me, so I assumed that you meant that I don't understand that fact (probably you still do).
Not at all Custos, terribly sorry if I gave you that impression.

Quoting Custos
I actually tried very hard to find systems that are complementing each other and also pairs that complement each other, so that I can achieve higher performances than if I would just use System A or B.
That's exactly the same approach I am using. Most traders like to diversify by trading different markets but personally I like to diversify by trading different non-correlated systems, especially trend and counter-trend systems.

Another approach is to follow 5 or 10 different unrelated trading systems, wait until a system reaches an abnormal drawdown and THEN trade its signals.
  • #386
  • Jul 27, 2011 3:50pm Jul 27, 2011 3:50pm
Quoting Mr Breakout

Another approach is to follow 5 or 10 different unrelated trading systems, wait until a system reaches an abnormal drawdown and THEN trade its signals.
I used to do similar, theory being from that point you have a good chance at getting a relatively big move, as just getting in and waiting and seeing I couldn't bring myself to do, yes the few that pay out pay out well if you have the confidence to hold them cause you get so many losers due to the high risk of the setup it's hard to have that confidence.
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #387
  • Jul 27, 2011 4:15pm Jul 27, 2011 4:15pm
Quoting Turveyd
if you have the confidence to hold them
In trading, confidence will always come from your backtesting homework.

For instance if your backtest reveals that your trading system had been averaging 200 pips a month per lot per currency for the last 15 years I can assure you that you will have all the confidence in the world to trade each signal without any hesitation or doubt of any kind.
  • #388
  • Jul 27, 2011 4:33pm Jul 27, 2011 4:33pm
Quoting Mr Breakout
In trading, confidence will always come from your backtesting homework.

For instance if your backtest reveals that your trading system had been averaging 200 pips a month per lot per currency for the last 15 years I can assure you that you will have all the confidence in the world to trade each signal without any hesitation or doubt of any kind.

But it's very hard not to cheat back testing just ignore a few you wouldn't take and take a few you would of missed and unless your back testing on tick data even scrolling through M15 isn't close to live trading and it's hard to factor in the time and everything else that goes into trading.

And realistically it'll take Months to accurately back test 15years of data, haven't got the time for that / patience.

I just check my edges hold up as edges and generally test with 6months of data, then putting my multiple edges together live is the issue left.
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #389
  • Jul 27, 2011 4:41pm Jul 27, 2011 4:41pm
Quoting Custos
okay, you probably have no knowledge in portfolio management. I try to reduce my drawdowns by employing different systems and applying them to different pairs. The systems and pairs complement each other. That way I can tune up my lot-size and achieve the 50 - 100% return.

I think it is you who has a hard time understanding that in portfolio management 1+1 is not always 2, but with the right combination it can outdo individual performances.
In portfolio management 1+1 is upper bounded by sqrt(2).
  • #390
  • Jul 27, 2011 5:05pm Jul 27, 2011 5:05pm
Quoting Turveyd
And realistically it'll take Months to accurately back test 15years of data, haven't got the time for that / patience.
You need of course a backtesting software like Omega Trade Station to test your trading ideas with 15 years of data, but the test itself could just take a few seconds/minutes to complete.

And once you start testing trading systems you will quickly notice that a lot of trading ideas (especially the seemingly "brillant" ones) simply do not work.

Unfortunately a lot of traders do not even bother testing their system, they just trade it immediately, hoping to make a killing. When they reach an 80% drawdown THEN they start to realize that their system has absolutely no predictive power and is no better than a random system.
  • #391
  • Jul 27, 2011 5:12pm Jul 27, 2011 5:12pm
Quoting Mr Breakout
You need of course a backtesting software like Omega Trade Station to test your trading ideas with 15 years of data, but the test itself could just take a few seconds to complete.

And once you start testing trading systems you will quickly notice that a lot of trading ideas (especially the seemingly "brillant" ones) simply do not work.

Unfortunately a lot of traders do not even bother testing their system, they just trade it immediately, hoping to make a killing. When they reach an 80% drawdown THEN they start to realize that their system has...

I'm a programmer and wasted 1 year writing EA's so I've back tested stuff, as tested I test my edge and test it every week on every pair I trade ( 14 currently ), I pull the pairs it's not favouring for a week until they come back in again, a system could work for 13 years of the 15 and have 2 years of constant draw down, but still show up as profitable, I like to reduce that.

Back testing, is no substitute for forward live testing, just don't do it with money you can't afford to lose, hence $170 account or 1/2 a days wages
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #392
  • Jul 27, 2011 5:43pm Jul 27, 2011 5:43pm
Example, last week M15 600Lwma Envelope 0.10% was my expected market turn back with trend zone, this week I'm noticing it's not quite getting there, so I'm not getting an entry, speeded it upto 540LwMA much better hit ratio on almost all pairs on the week.


See I live tune my system based on recent performance, next week it might just dip to the other side which is the SL zone so might be getting in too early, if so back to 600 or whatever they are using, I merely follow them.

That is our mission statement, do what the big players do follow them no need for getting all creative.
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #393
  • Jul 27, 2011 5:58pm Jul 27, 2011 5:58pm
Quoting Turveyd
See I live tune my system based on recent performance
Be careful Turveyd, a robust trading system must work over an extended period of time and with the same exact trading rules.

If a trader needs to frequently "fine-tune" his system depending on recent market conditions then "Houston we have a problem!"

PS: What programming language do you use, C++ ?
  • #394
  • Jul 27, 2011 6:03pm Jul 27, 2011 6:03pm
Quoting Mr Breakout
Be careful Turveyd, a robust trading system must work over an extended period of time and with the same exact trading rules.

If a trader needs to frequently "fine-tune" his system depending on recent market conditions then "Houston we have a problem!"

PS: What programming language do you use, C++ ?

Thats an opinion and 1 I no longer subscribe to, I got for continually adapting.

I do C++ and C# when I have to, more VB6 VB.NET Access / Excel apps faster Dev Times than C stuff.
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #395
  • Jul 27, 2011 6:12pm Jul 27, 2011 6:12pm
Quoting Turveyd
Thats an opinion and 1 I no longer subscribe to, I got for continually adapting.
You are of course free to experiment with any trading idea and strategy, but if you keep changing your trading rules then you are curve-fitting your system, unless you are simply simply trying to determine the right values and parameters for your system.
  • #396
  • Jul 27, 2011 6:18pm Jul 27, 2011 6:18pm
Quoting Mr Breakout
You are of course free to experiment with any trading idea and strategy, but if you keep changing your trading rules then you are curve-fitting your system, unless you are simply simply trying to determine the right values and parameters for your system.
Constantly curve fitting to best handle different market conditions is the plan and works, obviously I don't know if it will be better for the next day but the odds are most of the time it will be.

I do not believe a Rigid System will work long term, adapt and over come!!
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #397
  • Jul 27, 2011 6:50pm Jul 27, 2011 6:50pm
Quoting Turveyd
Constantly curve fitting to best handle different market conditions is the plan and works, obviously I don't know if it will be better for the next day but the odds are most of the time it will be.

I do not believe a Rigid System will work long term, adapt and over come!!
I respect your opinion but let me ask you a question Turveyd : Do you really think that markets change over time?

Look at the daily currency charts from 1980 to 1990, then from 1990 to 2000, then from 2000 to 2010 and tell me if you see ANY difference at all. You won't find any because the market can only do 3 things and 3 things only : go up, go down, go sideways. That's it, go up, go down or go sideways. And if the trader can capture these up and down movements he will make money, simple as that, there is no need to "adapt" to anything.

In fact, if your trading system is using volatility stops/trailing stops then it is already "adapting" to the market.
  • #398
  • Jul 27, 2011 10:38pm Jul 27, 2011 10:38pm
Quoting Mr Breakout
I respect your opinion but let me ask you a question Turveyd : Do you really think that markets change over time?

Look at the daily currency charts from 1980 to 1990, then from 1990 to 2000, then from 2000 to 2010 and tell me if you see ANY difference at all. You won't find any because the market can only do 3 things and 3 things only : go up, go down, go sideways. That's it, go up, go down or go sideways. And if the trader can capture these up and down movements he will make money, simple as that, there is no need to "adapt" to anything.

In...
Knowing that the market goes up and down is the easy part. You can look at a chart in hindsight and instantly see how the market moves up and down. What takes years to master is figuring out why a market STOPS going up and why a market STOPS going down. In other words, it is extremely easy to get into a market, but it takes a stroke of genius to know where the exits are in order to extract maximum profit.
  • #399
  • Jul 27, 2011 11:30pm Jul 27, 2011 11:30pm
Quoting Troikaone1
What takes years to master is figuring out why a market STOPS going up and why a market STOPS going down.
Granted, it takes years to master the art of chart reading.

However, WHY is the market is doing this or that is absolutely unimportant, what is truly important is to time its up and down movement.

Quoting Troikaone1
In other words, it is extremely easy to get into a market, but it takes a stroke of genius to know where the exits are in order to extract maximum profit .
That my friend is the role of the backtest, only your backtest can show you the optimal point of exit for your particular strategy.

Here is a simple and possible exit strategy for example: Exit your long (short) position if the price does not make a new high (low) after 15 bars. Or exit your long (short) position if the price closes below (above) its simple 100-bar moving average.

See, no need to reinvent the wheel, the trader can test hundreds of simple exit rules like that and find the optimal one for his particular system
  • #400
  • Jul 27, 2011 11:35pm Jul 27, 2011 11:35pm
Quoting Troikaone1
Knowing that the market goes up and down is the easy part. You can look at a chart in hindsight and instantly see how the market moves up and down. What takes years to master is figuring out why a market STOPS going up and why a market STOPS going down. In other words, it is extremely easy to get into a market, but it takes a stroke of genius to know where the exits are in order to extract maximum profit.
Rather than try to get Max profit on every trade...shoot for something reasonable, like 80%.

Whatever the system, it only has to have a profit factor > 1 to be profitable
2 traders viewing now