Robot Lab Project

So fare 2 moving average related systems are covered.

Next we'll move to some Grid based systems.

I will start with a system created and fully explained in another thread (The Ladder System).

Of Course. You can use SL_Pips Parameter (Depending on how many digits your broker provide:if 4 digits then 50 means 50 pips, if 5 digits then 500 means 50 pips.)

Guys, please note that for newer posted versions of any Robot ONLY the new or changed Parameters are explained in that post because old parameters would have been explained in earlier version's posts.

GIRD Systems
Ladder_System_01

Robot Attached.

Input Parameters :

1. Fixed Ladder Step Size in Pips,
2. Unified Stop Loss in Pips,
3. Maximum Number of Open Orders,
4. Fixed Lot size for all orders.
5. Lot Size as percentage of equity (added for this version, although not recommended)

Sample Test Results for EUR/USD April 1st,2014 till today

If you read the description of this system in its own thread, the robot will show many consecutive small losses followed by a big win and so on.

ABOUT ROBOTS

I would like to highlight some issues regarding working with robots on live accounts.

 

1. Use small position sizes, specially for robots that open multiple positions
2. Run multiple robots on same account, preferably with different systems, or at least different parameters or time frames
3. Run a robot against 2 or more instruments, preferably not highly co-related (positively or negatively)
4. Run robots on different brokers
5. Once account balance reaches certain multiples of initial investment(2x, 5x, 10x ... etc), WITHDRAW at least 50%.
6. Part of Withdrawn money can be used to open a new account at a different broker or initiate a new robot.

What I mean is that Robots can AUTOMATE your trading, BUT you need to MANAGE your Money, your Accounts AND your Robots as well.

Very interesting. Keep doing nice work,

May I ask how long have you been programming with mql language?

Thank you.

You're welcome. A year plus.

Ladder_System_01

Few words about the Ladder System:

1. The robot uses an implied method for opening the first trade. Simply for the applied time frame, if current bar is up it will go Long, if its down it will go Short (on market price). After successfully opening the first trade it will create Stop Orders of same type, spaced by the Ladder_Step_Pips. Open orders (Market + Pending) are equal to the Max_Open_Orders all with same lot size either Fixed_Lot_Size or variable (Lot_as_EQ_Perc), and all have Unified_Stop_Pips.
2. Stop Loss will is trailed by the Unified_Stop_Pips. So always all market orders will have the same stop loss.
3. Market Orders will be closed either when stop is hit, or when ultimate Take Profit is reached.
4. Ultimate Take Profit for all market orders will be one step after the last order.
5. User can control the first order, by attaching the Robot to the chart when the current bar is Bullish if he wants to gol long, or Bearish if he wants to go short.
6. The best use of the ladder system would be at anticipated tops or bottoms after price retraces. Where risk will be at the minimum and rewards can be huge. OR at main support/resistance areas where a pull back, and main trend continuation is anticipated.

Happy Trading,

Ladder_System_01

Chart Example explaining the above post

Attached Image (click to enlarge)

ROBOTS USES

 

1. System Back Testing:
   1. You might learn about a profitable trading strategy from a book, a forum , a friend. So what to do next ?
   2. One way is to develop a robot based on that strategy and apply it to different periods, and see how profitable it is against historical data
   3. Calculate the Expectancy ratio of the Robot. ER = [(Average Profit x Win Percentage) - (Average Loss x Loss Percentage)] / Average Loss
   4. Positive Expectancy ratio indicates a profitable strategy. ER is meaningful if calculated over at least 30 trades, the more the better.
   5. As an Example, if we calculate the Expectancy ratio for Ladder_System_01 as posted in Post # 24 (which has 53 trades):
      1. Average Profit (384.58) x Win Percentage (45.28%) = 174.14
      2. Average Loss (103.25) x Loss Percentage (54.72%) = 56.50
      3. ER = (208.75 - 56.50) / 103 = 1.14
      4. So what does this mean ? It means that you can gain 1.14 Dollars for each Dollar risked.

      
      

   6. Expectancy ratio is used to compare different systems (robots). The higher ER the better system.

   
   

Expectancy ratios for previously posted system results :




Note : MA_CROSS_WT_02 is not included because only optimization results were posted.

ROBOTS USES

Trading can be a highly emotional activity which can lead to failure.
Discipline is an important factor in trading.

Now, where you can find a more discipline and emotion-free trader than a Robot.

Next on my development list is :

N-Bars Breakout Entry X-Bars Retrace Exit

This system is some how similar to Turtle Trading System.

Input Parameters:

1. N_Bars_BO : N-Bars Break Out (for Entry)
2. X_Bars_RT : X-Bars Retrace (for Exit).
3. Fixed_Lot_Size : Fixed Lot Size
4. Lot_as_EQ_Perc : Lot Size calculated as a percentage of available Equity

​Example Chart below shows the system with N = 20, and X=10

N-Bars Breakout Entry X-Bars Retrace Exit

Robot Attached

Back Test Example on EUR/USD Daily from 2010 till today

Thanks a lot.
This one is the most interesting out of all so far.

All posted robots so far are Trend Following Robots, The more interesting Robots are still in my Pockets (or shall I say in my head ... LOL).

Stay Tuned.

Not mean to discourage you but make sure that your tick data is 100% accurate on 1M level to get real backtesting results.

Data offered by brokers/Metaquotes is far from complete and making a live robot based on it usually results in serious losses.

No worries here. Just testing the ideas. Will do demo first of course.
The last one for example is interesting for me not because it is especially profitable but because it is so simple and I don't think I came across this kind of robot before.

Good. For example forextester offers historic datas;
http://www.forextester.com/HistoryServiceDescription

They can be imported to for example in MT4 with a tool, at least used to be.

I can guarantee testing results will be completely different with accurate history data.

MA crosses need a better data set to be tested on IMO looking at USDJPY you can very easily curve fit a trending system to the last couple years or so . To give a true test of a systems robustness 7-10 years test are sufficient. IMHO happy pipin'