Moving average crossover

[meanreversion]

One of the simplest trading strategies is the moving average crossover. Now, they say simplify to simply fly, and you can't get any simpler than this puppy. I've been testing it out in EUR/JPY and the last few weeks have been very profitable. However, in USD/CAD it's sucked a little as there are too many "chops". 1) Does anyone use this approach and 2) what worthwhile filters can be applied to avoid those tight, choppy markets?

 

[tenbobtrader]

timsk uses this method I believe

 

[tenbobtrader]

some stuff on

MA SMA EMA tools and filters here for anyone interested

errrrr no in answer to your question

Decision Point ®: Learning Center

Andy

 

[GladiatorX]

I use the Moving Average crossover ONLY when price is a certain distance away from the 100EMA... I do this using price oscilator (1MA short moving average/100MA long moving average) therefore it tells me the distance price currently is from the 100EMA... When it is a certain pre-determined distance, i will wait for crossover its the opposite direction and play them to the opposite extreme of the 100EMA...
When the market is chopping, i don't enter because price usually on the MA, but since i don't want to miss the crossover that occurs before a big breakout from the consolidation, i enter on the breakout of the current high/low.

Hope that helps...
I've heard people use ADX filter... Use their eyes... Only trade when price is near the extremes of a Moving average channel...

 

[meanreversion]

I find ADX to be pointlessly lagging sometimes. I think maybe look at one time frame higher and guage how choppy/trending the market is. Thanks for this input guys

 

[GladiatorX]

I find ADX to be pointlessly lagging sometimes. I think maybe look at one time frame higher and guage how choppy/trending the market is. Thanks for this input guys

Personally never used it so i don't know much... It depends if you are trying to catch trend continuations or reversals ... By looking at a higher-time frame and waiting for it to trend, you've missed the crossover that caused the big move...

However if you use say use an MA to define Higher time frame direction, then use lower-time frame to enter on pullbacks and then confirmed continuation, that could work well.

If you included an oscilator i think it can work well in ensuring you don't enter when the market is 'in the middle', so you only enter when it is near the extreme of an oscilator... Do this in conjunction with a higher-time frame trend and it could be an excellent way to enter the market on pullbacks within an existing uptrend.

 

[meanreversion]

I've been using MA in daily time frame to determine direction along with RSI on 4hr time frame to wait for pullback, pretty much exactly what you have described.

However, it means sitting out some of the big reversals, as they are not "trend".. e.g. EUR/JPY goes down 500 pips then rallies 500 pips..... I miss out on the huge rally as I'm simply waiting to short it again. Those are the kind of moves I am looking to catch (if I can get 200 out of 500 I wouldn't complain).....

 

[meanreversion]

Michael Pryor details this approach in his book on spreadbetting strategies.

 

[GladiatorX]

Yeah thats that i'm saying, through that methodology - You will only catch existing, already occured trends, rather than that start or reversals...

Try putting a crossover (that are relatively far apart) 20 and 50 for example and then placing a moving average channel on price that is relalatively wide based on previous data... Enter on crossover and hold until price closes above channel... Can work very well as a target.

I would never use a MA crossover as an exit signal... I think thats an important part of this strategy that makes many fail... The MA can remain flat during choppiness but price gives 200-300 pip opportunities... Its a bad exit strategy.

 

[GladiatorX]

Pryor

Michael Pryor details this approach in his book on spreadbetting strategies.

I'll read up on that now...

 

[meanreversion]

The problem with using MA as an exit strategy is it gives back too much money. You need to overlay a money management technique to improve this, and if nothing else it helps stay sane. For example, s/l at 2 ATR and t/p at 6 ATR. Once the position is stopped/profit is taken, no position is re-established until the MA crosses over again. This way you don't have the stress of always being in the market.. and you're also unlikely to see massive profits given back.

How do you construct a "moving average channel"?

 

[GladiatorX]

If you placed like an 100EMA - 20EMA and 50EMA... Wait for an uptrend...

20EMA - 50EMA - 100EMA...

Wait for 20EMA to close below 50EMA... Wait for 20EMA to close back above 50EMA - It seems like a very good entry into trending markets...

OR
Wait for 20EMA to close below 100EMA to trade the reversal in the opposite direction to the trend...

You get in trade reversals and continuation, however NOT the pullback from the downtrend - We could enter this using an hourly setup.

 

[GladiatorX]

The problem with using MA as an exit strategy is it gives back too much money. You need to overlay a money management technique to improve this, and if nothing else it helps stay sane. For example, s/l at 2 ATR and t/p at 6 ATR. Once the position is stopped/profit is taken, no position is re-established until the MA crosses over again. This way you don't have the stress of always being in the market.. and you're also unlikely to see massive profits given back.

How do you construct a "moving average channel"?

Do you want to develop this over Skype - I'm making a trade journal on using indicators so this could be a good thing to develop, demo and backtest.

I think the most important thing is trade management here, the entry is good enough to get you into some nice positions.

 

[Ingot54]

Fyi

The problem with using MA as an exit strategy is it gives back too much money. You need to overlay a money management technique to improve this, and if nothing else it helps stay sane. For example, s/l at 2 ATR and t/p at 6 ATR. Once the position is stopped/profit is taken, no position is re-established until the MA crosses over again. This way you don't have the stress of always being in the market.. and you're also unlikely to see massive profits given back.

How do you construct a "moving average channel"?

Very good thread developing here guys.

Could I direct you to this link

New profitable System - Sidus @ Forex Factory

where you will find details of "The Sidus Method".

The author began with his model which he said was working ok for him. But as things evolved, a couple of adjustements occurred which seemed to make the method more profitable. Eventually "Sidus V.2" was born, and can be downloaded as a .pdf file, as can V.1 original method.

I would like to direct your attention to a post within that thread - #386 by OrangeRoshan. This was an attempt to get out of trends without giving back too many pips - a common problem with MA cross-over methods. The poster also gives a free .pdf of his method for your interest.

Whipsaws are always going to be a problem - just trade the higher TF and take earlier profits. Then move on to the next trade.

Best wishes

Ivan

 

[meanreversion]

This method looks like a variant of the 3 MA system developed back in the 70's. In essence, you add an additional MA to get further confirmation. There is an elegance and simplicity here which is appealing, although there needs to be an additional filter to avoid tight, choppy markets. Or alternatively, one could use the MA crossover as confirmation for another signal. The main drawback on the MA is of course the lag, but sometimes it's far better to miss out on the first 10-20 pct of the move and have confirmation, rather than trying to pick tops and bottoms.

 

[wackypete2]

Try using MACD on a larger time frame as a filter. If the MACD is signaling short then only take the short entries on the smaller time frame, if the MACD is signaling long then take only the long entries. As an exit strategy try using a shorter or midpoint ma (NOT a crossover). If price moves through the shorter ma by a set # of pips then exit. I sometimes use variations of this method for scalping during news trending days. Should work for longer term trading using higher time frames, but I never used it that way.

Peter

 

[meanreversion]

Ah, finally someone who uses a simliar method to mine. I use MACD on standard daily settings (12 26 9) to let me know which way the "trend" is. I normally trade off 4hr RSI.. I find that daily MACD rarely crosses zero (I'm talking about MACD line itself, not signal or histogram) but when it does, it's usually significant. So, for the times when it's below zero look for chances to go short, and above zero, look for chances to go long. Unfortunately there is only one filter which prevents all whipsaws ("don't trade") so this method is as prone to losses as any other.

 

[Liquid validity]

I know its an old thread but been working on something similar myself.
Best filter I've found so far is to simply switch to S&R breakout after 3 losing trades - without fail, every time no matter what. I use an sms alert system cause I hate screen watching and babysitting intra trends.

Doesn't always come good, but on the whole its salvaged some crappy days, either minimising loss or turning a small profit. Mainly run it on EU as an intra strat, 0630-1100 & 1300-1530 GMT (cba to catch U.S close )
Set sms price alerts for the 1100-1300 window just in case, usually time for a break though Nice & simple works for me.

Also no reason not to tune the dual/triple MA to suit the typical volatility of different pairs.

 

[NVP]

hey all

I like combining the 20 sma and 80 sma on a chart and trading the real price (or a very very low sma) when its above or below both..........could add a third 50 sma to be more fussy and also layer in some price action only going in when the price retraces and then rebreaches/diverges further away from the sma's

like this....I need the previous price (to signal the breach signal) to also be above/below the 20 50 80 as well

N

 

[NVP]

I am also testing convergence ideas on exteme 500 sma divergence values

N