stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

Macd Stochastic Fractal MA5/25 Daily Marylin
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=1448
Page 137 of 141
Author:  milanese [ Wed Jan 22, 2014 2:12 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

mirador » Wed Jan 22, 2014 1:19 am wrote:Here are better settings for EURUSD in .set file
AUDUSD coming soon, GBPUSD will be next

These are more agressive settings. DD is between 35% and 40%.

Just beware, backtesting results vary from a broker to an other, so optimized settings may vary too.

It's not easy to optimized because changing one parameter could affect other already optimized. I need to better understand all the parameters. There are so many.

UseStochH4ToTakeTrade activate StochH4LevelMax
UseTakeProfitOn3Macd activate DeltaPipsInterMacdToClose
UseStochToConfirmCloseOn3Macd also related to DeltaPipsInterMacdToClose

Am I right ? Are they other related parameters that I should know ?

Thanks
You will find all parameters, the EA uses, in the UserGuide attached at the first post explained.
But beware thinking with the optimizer function from Empty4 backtester you could create settings working in real trading, you will fail..

Cheers :)

Tommaso
Author:  SpiderX [ Wed Jan 22, 2014 2:28 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

Hi Tommaso,

If i am not mistaken, he is using tick data simulation on tester, so maybe there is some hope.
Probably his backtest takes an exceptional amount of time given that it is backtesting using ticks.
Maybe mirador can share how long it takes to backtest MM ?

Cheers
Author:  mirador [ Wed Jan 22, 2014 5:21 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

You will find all parameters, the EA uses, in the UserGuide attached at the first post explained.
But beware thinking with the optimizer function from Empty4 backtester you could create settings working in real trading, you will fail..

Cheers :)

Tommaso
Thanks Tomasso I will read it again...

I agree, we must be very carefull with Empty4 backtester. The problem is the actual settings in the .pdf does not give positive results for most pairs on demo account neither. I know that we need to be patient with this EA that is waiting for big wave. Marilyn has great potential, Phil proved it in 2013. Not enough work has been done on settings on our side.
Author:  mirador [ Wed Jan 22, 2014 5:41 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

Hi Mirador,

35%-45% DD seems a bit high...it seems to suggest your risk setting is too high.
Have you checked what is the reason for the high DD ?

Cheers
Yes it's a too high. I did not concentrate on the reason yet, neither on setting to reduce it. But from what I see, MM needs a big trend to grab big profit. It can wait many months taking a serie of small lost. When there is no clear trend, Marilyn is not good. She needs action.
Author:  mirador [ Wed Jan 22, 2014 5:43 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

SpiderX » Tue Jan 21, 2014 9:28 pm wrote:Hi Tommaso,

If i am not mistaken, he is using tick data simulation on tester, so maybe there is some hope.
Probably his backtest takes an exceptional amount of time given that it is backtesting using ticks.
Maybe mirador can share how long it takes to backtest MM ?

Cheers
It takes about 20 - 30 minutes each run (6 years of tickdata). Many many runs are needed.

Cheers
Author:  hardpike [ Wed Jan 22, 2014 1:53 pm ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

One of the risks is "over-optimization" this means that you can have a looovely backtesting from an optimized period of time (lets say 2009 end of 2013) and then 2014 fail....

one way to check this is runing an optimization from 2008-2012 for example

and then run 2013 (like a new year live)

if the optimized settings survive, that is a winner....
Author:  mirador [ Thu Jan 23, 2014 11:43 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

Here are new settings for AUDUSD.

UseTakeProfitOn3Macd is a powerfull parameter to reduce DD. With MM_AUDUSD2.set the only change is that parameter ON. DD goes from 29% with set 1 to 11% with set 2.

If you do the same change with EURUSD settings, DD will go from 34% to 25%.

No surprise if I tell you that the profit factor is less but, as you will see, it's still very good.
Author:  mirador [ Thu Jan 23, 2014 11:55 am ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

Sometimes, I have this message when running Empty4 strategy tester:
"Alert: SS_SupportResistance-Indicator won't work, less then 100 bars found"

Strange because I did not change any backtest period, TF nor SRZone parameter. I've tried many things, without success. I was wondering if it happens with live account.

Thanks for your help
Author:  milanese [ Thu Jan 23, 2014 12:04 pm ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

mirador » Thu Jan 23, 2014 11:55 am wrote:Sometimes, I have this message when running Empty4 strategy tester:
"Alert: SS_SupportResistance-Indicator won't work, less then 100 bars found"

Strange because I did not change any backtest period, TF nor SRZone parameter. I've tried many things, without success. I was wondering if it happens with live account.

Thanks for your help
as I had written before this means you don't have the bars needed for using the indicator, it needs at least 100 bars in past for calculation, better is more, so your test is worthless..

Cheers :)

Tommaso
Author:  milanese [ Thu Jan 23, 2014 12:08 pm ]
Post subject:  Re: Macd Stochastic Fractal MA5/25 Daily Marylin

zanzistor » Thu Jan 23, 2014 12:04 pm wrote:
mirador » Thu Jan 23, 2014 11:55 am wrote:Sometimes, I have this message when running Empty4 strategy tester:
"Alert: SS_SupportResistance-Indicator won't work, less then 100 bars found"

Strange because I did not change any backtest period, TF nor SRZone parameter. I've tried many things, without success. I was wondering if it happens with live account.

Thanks for your help
Mirador, try to start the backtest 4 months later...
Example
you want a from 2008 01 01 starting...

You will need feed from 2007 06 01

The error will disapper...

Thanks for your job!!!! Really appreciate
you should have at least a feed starting one or two years before.. it is needed for the SR calculations

Cheers :)

Tommaso
All times are UTC Page 137 of 141