PipMaker v1 - Price action based EA - page 2

 

TF matters

don_forex:
I will start posting settings that will pass a three year backtest. I ask that people do the same since there is more of you than just me. Maybe somebody will hit a great combination. Any pair. TimeFrame doesn't matter. I use the 1M simply because I haven't converted to all the other TFs. Let's see what we can come up with... Don

I think that TF matters because you open trades at the beginning of the candle, so the highest TF the less trades.

 

...and when you start?

Just from my initial testing and optimizations on EURUSD and GBPUSD, it doesn't matter when you start (at a high or low of the year is worst case), you will usually always end up with drawdowns of ~50-60% of equity. As long as you can stomach that, this grid system can make tons of money.

Carl

 

Great job and thanks for sharing don!

I was looking for something like this for some time now.

When I saw "Price action bases" from you, I knew it would be good stuff.

 
crodzilla:
Just from my initial testing and optimizations on EURUSD and GBPUSD, it doesn't matter when you start (at a high or low of the year is worst case), you will usually always end up with drawdowns of ~50-60% of equity. As long as you can stomach that, this grid system can make tons of money. Carl

Second that Carl.

I am prepared to look at that the first year.

Break-even second year and start making money third year and let the moneymachine run while relaxing

Unless something very weird happens to the forex market or your broker goes broke

 

Worst Case Scenarios...

Ok... I wanted to put this grid system at the biggest disadvantage I could. So, I picked one of the worst times to start the grid for the GBPUSD. I picked Nov 27, 2005. There was an intermidate term bottom here, and price rises swiftly from here through today. With a 1 pip step using 0.01 microlots, 15min timeframe and $10 target, the balance in May, 2006 is over $15K... however your equity is $0. Bust. I have tried a lot of different values of step, lots and profit targets for the GBPUSD on the 15 min timeframe. I don't think its possible. That timeframe sucks for the GBPUSD.

I have found that the larger timeframes will alleviate the problem of overpositioning on one side. However, if there was a better way to wash the portfolio of the losing equity side, you would be able to survive long trends. The hope is the currency will churn for a while... enough to rid yourself of some of the losing trades. This EA does a decent job, but once those positions are deep on the losing side... they stay losers and kill your equity. Especially when the GBPUSD never looks back like it has.

Volatility is a beautiful thing with this EA. Trends are fine, as long as there are some decent retracements.

Just for fun, i'm going to try the 5min timeframe...

Carl

 

Optimization tip...

When using the backtester, you don't need to use every tic or whatever, just use the close/open of the next bar. Not only does that make your optimizations/backtests much quicker, but that is how the EA trades anyway, so you don't lose any accuracy.

The above might be a no-brainer, but I was sitting here for hours looking at the tics go by... waiting. I'm so dense...

Carl

 

GBPUSD Optimization...

Lots = 0.01

Step = 1

Timeframe = 4H

Increment = 0.005

Profit Target = 200

Spacing = 1

From: Nov 2005 - Jun 2007

Starting Balance: $10,000

Ending Balance: $40,345

Lowest Margin Percent: 308%

------------------------------------------------------------

This optimization is done at a (intermediate) low for GBPUSD. So Prices never returned back to those lows. At these settings the lowest Margin was just over 300%. At todays prices, the Margin sits at 2400%. The EA has done a good job of eliminating the drawdown trades.

Carl

Files:
 

Sorry to be unavailable

My wife took my Sprint card to Florida for her little vacation. So, I can only get internet when I am in truck stops. I have been working on some revisions that will greatly enhance the performance of the EA.

You will notice the errors still do occur. I am working on that. I have noticed that only using the same type close routines the errors do not occur, i.e., positive buys and a losing buy, and positive sells and a losing sell. I need help in isolating the lowest and highest orders. When the errors occur, multiple losing orders close when only one should. If anyone out there could help out, that would be great.

I am working on version 3 which incorporates a LWMA to determine trend, it only opens buys above the trend and sells below the trend. It still sets buys above and below the highest and lowest buys. Same for the sells. Still having trouble with the opposite closings. pos buys losing sell, pos sells losing buy...

I will try to get online later tonight.

Don

 

Clarification

Concerning the openings. there is an internal boolean variable called Boost. When it is set to true, it will set orders every spacing increment. So, in that respect, the EA does not only open orders on the open of the candles. I couldn't remember if I left it on or turned it off...

Don

EDIT - I looked at the code on the first post and it is turned off (false)... Turn it on by replacing the false with a true. In the next revision, I will make it an external variable.

 

Thanks...

Don,

Thanks for that info. I did not look close enough at the code to see that. What I did do was add a trend change filter. I've seen that the code will exit positions upon a profit target. What I did was add the "peak of ADX" filter which signals a possible trend change. So, by using the next higher time frame use the ADX to allow the code to then check for profit target and exit positions as nesessary. The results have been encouraging. I am still testing, but find that the delay to the end of trends (or consolidations), squeezes out more profit from some of those positions. So in essence, the profit target becomes minimum profit target.

Thanks for sharing this EA. I am infinetly interested in these type of EA's.

Carl

Reason: