Cost Averaging System - page 17

New comment
 

Open over the weekend

As a matter of interest here are the differences between closing time last week and opening time this week. The EAs were allowed to be open for the weekend.

Cost Averaging v3 lost $171.

Cost Averaging v4 lost $44.

Cost Averaging RSI gained $279.

I have taken detailed statements ASAP of both positions of each EA if needed.

John

 

Looking at what the system is doing and trying to understand how it works has got some things cleared up in my mind. I am working on the premise the cost averaging works. Also, comment smade by Aaragorn, about trend trading and Crodzilla about waiting for the peaks in ADX caught my attention. I started to think, "if I am looking for trend exhaustions and why not try to trade in the direction of the trend in the larger timeframe."

Basically, I want to buy the dips in an uptrend and sell the rallies in a downtrend, when I perceive that the shortterm selling/buying is over. That is best shown by indicators like RSI, that can point out oversold/overbought areas. Also, longer term trends are evident when prices close above Moving Averages. Based on these ideas, I modified the EA and will be sending out a revised copy this evening.

I know some users have good quality data. Please backtest these to get an idea what settings work. Again, we want the robust settings, not necessarily the most profitable settings.

Thanks everyone.

 

current results are attached

 
rarango:
I did a lot of back testing to find settings that do not blow account at some point.

Some of them give really good results in back test for aproximately two years with 90% accuracy.

For some pairs I could not find good settings: GBPUSD and GBPCHF

I am starting forward testing with this settings.

all pairs are in M15

Rarango and others,

Did you get a chance to backtest the new RSI with Trend EA? I would love to see what kind of settings are working for this EA.

Remember, in the default setting, the EA trades off RSI of 15 minute bars and only in the direction of the 240 minute bar Moving Average.

Maji

 
Maji:
Rarango and others,

Did you get a chance to backtest the new RSI with Trend EA? I would love to see what kind of settings are working for this EA.

Remember, in the default setting, the EA trades off RSI of 15 minute bars and only in the direction of the 240 minute bar Moving Average.

Maji

Maji, here are some optimized backtests over 2006. Focus on minimal drawdown.

EUR$ on 15M TF

and

$JPY on 5M TF

(It seems they both have difficulty with long trades.)

 

Ea

Maji I did PM you my email but I havent received the EA yet.

Please have a look at your PM's when you have a chance

Regards

Z

 

Z,

I can't remember about that PM. Sorry about that.

Please email me and I will send you the system.

Maji

 
HerbertH:
Maji, here are some optimized backtests over 2006. Focus on minimal drawdown.

EUR$ on 15M TF

and

$JPY on 5M TF

(It seems they both have difficulty with long trades.)

Hebert,

Thank you for your effort. I want to point out a couple of things.

First, you set the take profit to 999. Basically that means that the trades are closed out by the RSI or the the MA.

Second, your buy and sell Levels are very asymmetric and does not really signify overbought/oversold areas. I think it means that they are curve fit results. Look at the parameters and see if they are realistic. That is the first test.

Try to get symmetric values and don't worry about the best... concentrate on finding the most robust. That means the values are more or less stable when the parameters are changed by a few percentage of its initial value.

Thanks again.

 
Maji:
Hebert,

Thank you for your effort. I want to point out a couple of things.

First, you set the take profit to 999. Basically that means that the trades are closed out by the RSI or the the MA.

Second, your buy and sell Levels are very asymmetric and does not really signify overbought/oversold areas. I think it means that they are curve fit results. Look at the parameters and see if they are realistic. That is the first test.

Try to get symmetric values and don't worry about the best... concentrate on finding the most robust. That means the values are more or less stable when the parameters are changed by a few percentage of its initial value.

Thanks again.

Maji,

The TakeProfit flattens-out at the same level as the maximum value found. There is no gain when limiting it to any lower value.

I have wondered about the asymetric aspect as well and cannot seem to get any better than this. I will give it some other tries, but maybe someone else is lucky in finding it.

Regards

 
HerbertH:
Maji,

The TakeProfit flattens-out at the same level as the maximum value found. There is no gain when limiting it to any lower value.

I have wondered about the asymetric aspect as well and cannot seem to get any better than this. I will give it some other tries, but maybe someone else is lucky in finding it.

Regards

Herbert,

Don't forget to take a look at the code. There maybe a small bug there that is forcing that asymmetry. However, try to keep the buy and sell levels symmetrical... that will help ensure "non curve fit" to a large extent.

1...101112131415161718192021222324...62
New comment