I wanted to share a strategy that I learned from an acquaintance of mine. Actually I know a whole group of professional traders (guys who make their living from trading) who trade this method. I have not read of it anywhere else, but don't know how widely known the method is or who else to attribute it to.
I traded it manually for a while on demo but the account got messed up so I don't have my performance figures, sorry. It is a trend following MA Cross system with some unique money management/scaling in and out functionality.
I wanted to know what people thought of MA cross strategies before posting this one, hence my questions/poll here -- "Quick opinion: MA Cross strategies" http://www.stevehopwoodforex.com/phpBB3 ... ?f=6&t=162
It seemed like I was in friendly territory so here we go!
I was hoping Steve or George or Gary or ?? might try to make it automated. It is purely a mechanical strategy but does have some "trading style" options that can be set that give it slightly different trading behaviors (but still totally mechanical).
The default method is a 24/7 trader and so doing this manually can be a bit of a bear -- it must be checked at least every 4 hours, even at night. And that is one of the main reasons I am hoping for an EA. (I am USA Pacific time - GMT-8 - which is about the worst hours you can be at for trying to trade forex.)
This post is VERY long but that is not because the strategy is very complex. Please read it all the way through. I have provided a PDF version of the strategy description as well. It might be easier to read than the on the forum.
I am trying for 3 things:
1) to OVER communicate all that is required for the strategy -- I don't want whoever takes this up to keep having to ask questions about too many things or for me to have to send a bunch of posts/emails/PMs to them to say "oh yeah, I forgot to tell you these 3 other mission critical functions..." Don't want to anger the Steve!
2) This post is using a "trading strategies description" template that I hope will be useful to others as they try to describe new trading ideas. If folks like the use of this template I/we will make a blank version available to download. Thoughts on the template should go here: http://www.stevehopwoodforex.com/phpBB3 ... ?f=12&t=80
3) I am hoping someone will find this trading strategy interesting enough to help program an EA for it.
Basic Info
------------------------------------------------------------------------------
Strategy Name: Miller MA Cross (MMAC)
Don’t know who came up with the system originally and I have never heard of it anywhere else, so I will name it after the guy who taught it to me. If others know the origin we can change the name.
Strategy Style: Trend Follower with two trading style options
Favored Market Conditions: Trending, but handles a certain amount of ranging well.
Entry Signal style: MA Cross with TF agreement based unit size
Expected Trading Frequency:
Style 1 -- 2 TF Agreement = Potentially intraday.
Style 2 -- 3 TF Agreement = Potentially intraday and would be “always in.”
Strategy Expected Moves per trade: 10-500+ pips
Description
-----------------------------------------------------------------------------------------------
Overall Trade strategy Description: When TFs agree on trend, we have a high probability trade. Lower TF Crosses are early signals of higher TF Price Action. This strategy trades 1 unit of risk in the direction of the MA trend per time frame compared.
EDIT -- UPDATE 1-4-12, 5:30PM: Basically this strategy views each TF being compared as worth a +1/bullish or -1/Bearish score. The scores of each TF are added together to get a composite score. That composite score is the number of units of risk and the direction of the trade. For example: 3 bullish TFs = +3 composite score = 3 units of risk.
If we are using .01 lots as our "per unit of risk lotsize" then a score of +3 = going long .03 lots.
If, on the next check of the TFs, we have a composite score that changes to +1 that means not all the TF's agree on on the trend so we have less confidence in the trend. We would reduce our position size from (+3 score = 3 units of risk = .03 lots) to (+1 score = 1 unit of risk = .01 lots).
That is pretty much the whole model. When we have more confidence in the trend we increase our risk, when we have less confidence we reduce or eliminate our risk. Confidence is determined by comparing the trend on several different TFs. The max risk is the max number of TFs being compared if they are all bullish/bearish. (End Update)
(See Scaling in section for more info on how this works.)
Why it works: Follows trend. Aligns trades for higher probability of move. Higher profits when strong trend.
What is the "EDGE": By having TFs agree, aligns trades for higher probability of move. Higher profits when strong trend. Follows Price Action of “The Boss” higher TF as it dictates trends to lower TFs -- until a reversal. We can then follow the reversal for as long as it lasts. Rinse, repeat.
Money management plays a large part in the success of this system -- multiple units of risk allow us to increase earnings on stronger, longer lasting moves and have less or no risk during times of ranging or “market confusion”.
Strategy Details
---------------------------------------------------------------------------------------------
Time Frames: Default: Style 1 = 8h, 1D or Style 2 = 4h, 8h 1D. The 8h will have to be synthesised with the number of periods in the MA indicator on a 4h chart. Default uses 10/30 period MA cross on all TFs as trade signals. (But EA to have all TFs and MA periods as variables... Also option to use additional TFs... see Options section.)
Times to Trade: Default is a 24/7 system. Check if MA has Crossed on all TFs at opening of lowest TF‘s new bar. (Can be setup to be 24/5 or only certain hours... see options section. Testing will show if this has any impact.)
Pairs To Trade: Majors, any low-spread trending pair. If using with multiple pairs, watch out for high correlation. We may want to keep the EA a single pair trader for now though.
Money Management/ Risk Management:
Default: fixed lot size as a base unit of risk. Can use any size lot that matches your risk tolerance and account size.
There is one unit of risk per TF used. So if you are using 3 TFs you could have a max of 3 units of risk open at any one time.
The system can have some times of DD as the MA cross lags price or becuase of spread costs during choppy bits. However the MM of this system is supposed to overcome this weakness quickly and still be very profitable.
Note: It is the MM side of this that makes this system very interesting to me and not just another “X,Y Cross” method.
This is the heart of this system: Overall position size is determined by Time Frame Agreement or specifically: Multiple Time Frame Trend Agreement.
Stated another way: “We have more money invested when we have a higher confidence of the trend. If all of the TFs agree, we can be more confident of the trend and have a higher investment in the market at that time. When TFs disagree we have less confidence in the direction of the market and therefore exit entirely or have only a small amount of risk in the markets at that time.”
This has the added benefit of helping 1) smooth out the lag of the crosses, 2) bring in many more [/b]pips during times of longer runs up or down and 3) larger pips on stronger moves helps overcome any losses due to times of ranging or several MA Crosses happening in a row without the markets making a definitive move.
System uses Scaling in/out: Yes.
Here is how the process works.
FastMA over(on top of) SlowMA = Go Long = +1 (Think of it as a Long unit of risk) = +1 Bullish Position
SlowMA over(on top of) FastMA = Go short = -1 (Think of it as a Short unit of risk) = -1 Bearish Position
At any given time, each TF will be giving you a +1/Bullish or or -1/Bearish signal.
These signals are added up to give a total Bullish position or total Bearish position.
3 TF Examples (not all possible combinations):
4h 8h 1d
+1 +1 +1= +3 = Bullish/Long 3 units -- Strong trend, higher confidence, larger risk
4h 8h 1d
-1 -1 -1= -3 = Bearish/Short 3 units -- Strong trend, higher confidence, larger risk
4h 8h 1d
+1 +1 -1= +1 = Bullish/Long 1 Unit -- Less confidence in trend, smaller risk
4h 8h 1d
-1 -1 +1= -1 = Bearish/Short 1 Unit -- Less confidence in trend, smaller risk
Note: With 3 TFs you are always in the market with a minimum of 1 unit or the max of 3 units.
2 TF Example (not all possible combinations):
8h 1d
+1 +1= +2 Bullish/Long 2 units -- Strong trend, higher confidence, larger risk
8h 1dP
-1 -1= -2 Bearish/Short 2 units -- Strong trend, higher confidence, larger risk
8h 1d
-1 +1= 0 Positions -- Out of the market until Confidence in trend returns, no risk
IMPORTANT: This system can have full position reversals. If you are bullish +2 and on the next MA cross check get signals to be bearish -2 then you must be short *4* units. 2 to get to a neutral position and 2 more to get to the -2 short position. Reverse for short position reversal to long position. Always be thinking of the total position the signals say you should have and then what must be done to get there.
Order types: Uses market orders for all entries or position reversals at time of MA cross check
Stops Management: Original Strategy does not use stops, but I think it would be good to have the ability to have some emergency based and/or margin based, equity based, Max DD style stops. JIC. (Just In Case)
Indicators required plus any custom settings:
I am hoping the EA could be built around AllAverages so that different MA styles and periods can easily be used/tested.
Default settings are:
MA Style = SimpleMA
Fast MA = 10 Period
Slow MA = 30 Period
Default = All TFs use the same MA style and periods.
These can all be adjusted based on testing. For instance, I have a friend who has been testing this using a 4/11 cross instead of the 10/30. Seems to do well. I would probably be switching to an EMA instead of the SMA. Seems to not lag as badly in general. Time will tell.
Trade conditions:
See Scaling in section above and Options for Startup below to see description and more info
I believe all entries and exits are discussed in the sections listed above. If more info is need, please let me know.
Trade Entry Long
Trade Exit Long
Trade Entry Short
Trade Exit Long
Max Number of Trades conditions: None for now.
Between trade conditions: There are options when scaling in the first time, and needing to be aware if closing or reversing trades depending on #of TFs used. But this is discussed in the above text.
Optional trade filters: None for now.
Specific Options to include:
On the number of TFs to compare:
I would like the option of comparing up to 7 TFs (default is 2 or 3). This will still allow the default setups but let us test the effects of using more TF agreements.
Each of the 7 TFs should have the following options:
TF 1(etc): Any of Empty4’s default TFs
MA Type: Default is SMA, but could be any of AllAverages MA options
MA Fast Period: Default is 10
MA Slow Period: Default is 30
Lot size for this TF: Default is that all TFs use same lot size but would be interesting to test results if certain TFs had a higher lot size weighting than others. Remember that you could have as many as (if all the same) 7 units of risk open at any one time if using all 7 time frames.
On When to trade the Crosses:
There are 2 options for when to trade the crosses. Options are mutually exclusive, cannot have both true or both false.
TradeOnNewBar: True/False -- Default -- EA always checks the state of each TF for crosses at the opening of the lowest TF’s new bar. This is also the least CPU intensive as it only has to check for crosses at the new bar.
TradeOnCross: True/False -- This would cause updated trade signals at the actual moment of the cross. It would require continuous monitoring of all TFs while looking for a cross. Other trading behaviours remain the same -- add all the TFs bullish/bearish signals for a total long/short position and then trade to match that position.
On building into a position:
When the EA starts how do we want to get into a trade -- these settings affect the building of your position on startup. Only one of the 4 options can be true. The impact of your choice will not be huge if trading 24/7 because there will be few starts. But if one chooses to trade 24/5 (stopping for the weekend) or use some other Time to Trade settings or start and stop the EA often, it could have a greater effect.
ScaleInOnLowToHighTFCrosses: True/False -- Default behavior. Build into positions starting by waiting for lowest TF cross on new bar. Basically follow as if a 1 TF strategy until a cross on the next highest TF at the new bar of the lowest TF. Then continue as the normal 2 TF strategy and build your bullish/bearish position. Keep adding TFs to compare as the next highest TF crosses on the new bar of the lowest TF.
GetInOnNewBarLowTFCrossWithFullPosition: True/False -- Wait for Cross on new bar of lowest TF and then check all signals from all TFs and take that position to start.
GetInOnLowTFCrossWithFullPosition: True/False -- Wait for moment of Cross of lowest TF and then check all signals from all TFs and take that position to start. (Same as above except don’t wait for new bar.)
TakeFullPositionOnStartup: True/False -- Don’t wait for any crosses or new bars, simply read the bullish/bearish signals of all TFs and take that position to start.
On when to trade:
Needs options to control at least:
Trade 24/7?
Trade 24/5?
Trade XX day and YY hours through WW day and ZZ Hour?
Trade everyday from YY hours to ZZ hours.?
(Maybe George’s recent time modules?)
End of Strategy Description.
Please let me know what you think!!
Cheers and many pips,
Rick
