Bears'n'Bulls EA - V1.05g - in forward- & live-test

[cheva282]

Thats the life!! learning from the bad try to make it better!

[banzak]

There are news in the first post.
http://www.stevehopwoodforex.com/phpBB3 ... 933#p34933

[auvergnat]

There are news in the first post.
http://www.stevehopwoodforex.com/phpBB3 ... 933#p34933

Hi nice results. You don't post your EA at this stage ?
Tof

[banzak]

There are news in the first post.
http://www.stevehopwoodforex.com/phpBB3 ... 933#p34933

Hi nice results. You don't post your EA at this stage ?
Tof

No, i will wait for the results.

Regards,
Banzak

[crs247]

Wow, these are ridiculous results: Account doubled in two weeks with just 5 % DD. I'm curious how this beats in the long-term.
Can't wait to play with the new version anyway!

[hopfi2k]

Its very interesting to analyze the trades in myfxbook.

What I really like are the high hit rates for entries and exists (at least for EURUSD and GBPUSD). The bot is doing something right here

Can this EA by backtested?

Cheers,
Andy

Wow, these are ridiculous results: Account doubled in two weeks with just 5 % DD. I'm curious how this beats in the long-term.
Can't wait to play with the new version anyway!

[banzak]

Its very interesting to analyze the trades in myfxbook.

What I really like are the high hit rates for entries and exists (at least for EURUSD and GBPUSD). The bot is doing something right here

Can this EA by backtested?

Cheers,
Andy

Hi Andy,
yes this EA can be backtested, but at the moment we have only good results with one criminal. The same setfiles doesn't work with the other. I started 2 demos and the 2nd I stopped already.
So you have to backtest and make setfiles for each criminal.

Lets see how the EA performs on long-term, the last try went fine for 5 months and killed the account in 1 week.

This ea will also get in the next version a hard S/L for emergencies and a hard S/L for the close-routine.

At the moment we're playing with the risk settings, to get the optimum lotsize for each pair. Thats the reason why the ea opened @ july to much and later to less.. (5 lots / 0.10 lots)

When I have some time I will update the actual entry-/exit-logic in the 1st post, then you can all see on what base this ea works. I will also update the template and maybe take some screenshots.

Regards,
Banzak

[hopfi2k]

Its very interesting to analyze the trades in myfxbook.

What I really like are the high hit rates for entries and exists (at least for EURUSD and GBPUSD). The bot is doing something right here

Can this EA by backtested?

Cheers,
Andy

Hi Andy,
yes this EA can be backtested, but at the moment we have only good results with one criminal. The same setfiles doesn't work with the other. I started 2 demos and the 2nd I stopped already.
So you have to backtest and make setfiles for each criminal.

Lets see how the EA performs on long-term, the last try went fine for 5 months and killed the account in 1 week.

This ea will also get in the next version a hard S/L for emergencies and a hard S/L for the close-routine.

At the moment we're playing with the risk settings, to get the optimum lotsize for each pair. Thats the reason why the ea opened @ july to much and later to less.. (5 lots / 0.10 lots)

When I have some time I will update the actual entry-/exit-logic in the 1st post, then you can all see on what base this ea works. I will also update the template and maybe take some screenshots.

Regards,
Banzak

Banzak,

Thanks for your detailed reply.

I'm wondering why you have different backtesting results with different brokers. Aren't you using Dukascopy tick-data?

Investigating your forward-demo shows that you are using highly correlated pairs (GBPUSD/EURUSD/USDCAD) which can - especially during high impact news - lead to serious DD of not MC. Wouldn't it be better to use low correlated pairs? (e.g. EURUSD/CHFJPY/GBPCAD).

Hard S/L should be a standard in any EA. Better to write off a certain amount than risking a MC

If you need a hand with backtesting, just PM.

Cheers,
Andy

[banzak]

Thanks for your detailed reply.

You're welcome.

I'm wondering why you have different backtesting results with different brokers. Aren't you using Dukascopy tick-data?

I'm using Ducscopy tick-data, but there is a difference - I don't know why.

Investigating your forward-demo shows that you are using highly correlated pairs (GBPUSD/EURUSD/USDCAD) which can - especially during high impact news - lead to serious DD of not MC. Wouldn't it be better to use low correlated pairs? (e.g. EURUSD/CHFJPY/GBPCAD).

Sure, you can try to get some setfiles for that kind of pairs.

Hard S/L should be a standard in any EA. Better to write off a certain amount than risking a MC

At the moment I have a closing-routine of x% Drawdown, when the EA closes the basket, but for safety reasons there must be also a hard-S/L if the ea looses connection. The same with the close-routine.


Regards,
Banzak

[hopfi2k]

Thanks for your detailed reply.

You're welcome.

I'm wondering why you have different backtesting results with different brokers. Aren't you using Dukascopy tick-data?

I'm using Ducscopy tick-data, but there is a difference - I don't know why.

Investigating your forward-demo shows that you are using highly correlated pairs (GBPUSD/EURUSD/USDCAD) which can - especially during high impact news - lead to serious DD of not MC. Wouldn't it be better to use low correlated pairs? (e.g. EURUSD/CHFJPY/GBPCAD).

Sure, you can try to get some setfiles for that kind of pairs.

Hard S/L should be a standard in any EA. Better to write off a certain amount than risking a MC

At the moment I have a closing-routine of x% Drawdown, when the EA closes the basket, but for safety reasons there must be also a hard-S/L if the ea looses connection. The same with the close-routine.


Regards,
Banzak

Banzak,

Could you elaborate on the differences with Dukascopy tick-data? Are you using Tickstory? I'm doing a LOT of backtesting on a dedicated machine, and only have difference in cases of: different spreads, different lot steps, different min/max lots. and stop levels. What brokers are you having this differences with? Are the differences substantial or marginal?

Cheers,
Andy