Frank EA and 10p3 EAs - page 28

 

10p3v0.03 EA.

It was stop loss for NZDUSD. It is first stop loss case since January 2008 (I am forward testing 11 pairs). Initial stop loss in the settings is 160 pips.

It is the statement for NZDUSD in pips and in deposit currency for this week:

 

10p3v0.03 EA.

It is all kinds of the statements (in rar archive inside attached file).

Winrar to open attachments https://www.mql5.com/en/forum/173397/page11

 

Frank_mod EA visual statements updated.

Visual statement using this indicator:

And please find the other statements in attached file.

As to the EA/settings/pairs/timeframe - look at the first post of this thread.

Winrar to open attachments https://www.mql5.com/en/forum/173397/page11

 

10p3v0.03 EA visual statements updated.

Visual statement using this indicator:

This EA is still in profit and I am continuing forward testing on same account and same settings with the same 11 pairs H1 timeframe.

As to the EA/settings/pairs/timeframe - look at the first post of this thread.

 

Fabulous progress with this EA!

Has anyone compared demo results with live account?

Are the modifications sufficient to avoid high volatility and to allow successful trades?

 

We are having just 3 kind of account protection we may use for those 2 EAs now (Frank_mod EA and 10p3v0.03 EA):

- timefilter. 10p3v0.03 EA is trading between 18 - 8 in GMT night and Frank_mod EA is trading within 21 - 4 which is more safe of course;

- more higher timeframe is more safe;

- limit the number of the pairs.

By the way those 2 EAs are performing good now.

Anyway, I am making this thread https://www.mql5.com/en/forum/178788 just to find some general solusion for some market condition. Because many EAs are doing very good profit (about +1,000 pips each just for 1-2 days) but some EAs are losing. It is fully correlated with other elite section idea concerning creation of EA Switcher (we are having the thread in elite section about it).

 

Frank_mod has performed beautifully so far for 5 months. Assessing risk in changing market conditions as described by newdigital is extremely helpful.

It is difficult to assess the risks long-term. But we need to think long-term, because the realtively low yield does not easily allow the replacement of a lost account.

I did some backtests on Frank_mod. There was a problem incident with EURUSD in early February 2008: the EA went long on Feb 5, and could have margined out accounts on Feb 7 for more aggressive approaches.

GBPUSD seemed to have been OK during the same time period.

Conclusions:

(1) Rare events could destroy the account.

(2) Multiple pairs may buffer the account, as suggested by newdigital. EURUSD, GBPUSD, CHFJPY contribute similarly. EURCHF has only about 25% of return compared to main 3 pairs.

(3) Interestingly, the Ichimoku indicated a trend change on Feb 4 for EURUSD on H1. Using a trend filter may decrease the risks, but will also decrease the returns (remember that using MACD in 10p3 was not the better solution). An alternative may be using a filter to close positions when it looks very dangerous, or put the filter in place temporarily only until the trend change is over, and then enter again as usual.

Has anyone else done backtests or compared with performance in live accounts?

 
Tokarus:
I have modified Frank_Mod EA further, to allow it to open new positions only when the Free Margin is at least 75% of the account Balance. This way we should be limiting the number of open orders.

I made it a parameter [called MinMargin] so you can play with various settings.

Cheers, Kris

Thanks a lot.

I updated link to your post here https://www.mql5.com/en/forum/176044/page2 and updated this thread as well What is new weekly in public and elite.

 

Well, my tests indicate that in trending markets the Time filter is more effective then my Margin filter.

The Time filter lets us stay out of the market and resume placing orders at specified time, while Margin fileter stops taking additional trades all together. This might not be that desirable, as cost averaging will become ineffective.

However, in a sideways market, my Margin filter should work better then time filter.

Cheers, Kris

newdigital:
Thanks a lot. I updated link to your post here https://www.mql5.com/en/forum/176044/page2 and updated this thread as well What is new weekly in public and elite.
 

Thanks for clarification.

I updated first post of the thread.