Look at this nice profit curve !!!!!!!!!

New comment
 

Guys,

What do you think about think this profit curve?

It kinda flat so we just need to increase lots to get more profit...

This curve is EURUSD since 2000, my EA is based on the H1

Please comment! what would be the risk in the future of such an profit curve?

Thanks in advance.

Best regards

Patrick

 
Demeuterpatrick wrote >>

Guys,

What do you think about think this profit curve?

It kinda flat so we just need to increase lots to get more profit...

This curve is EURUSD since 2000, my EA is based on the H1

Please comment! what would be the risk in the future of such an profit curve?

Thanks in advance.

Best regards

Patrick

It looks good Patrick, but the curve on its own is useless without knowing whether you are showing any out of sample results. Other items from the strategy tester results such as profit factor, average profitable trade and average losing trade are also useful in assessing the potential that it may be a profitable EA.

Also, does the EA make decisions based on current bar or last closed bar? This is then very important when linked to whether you tested on the 90% every tick model.

Paul

http://paulsfxrandomwalk.blogspot.com/

 

Thanks Paul,

My orders are based on Close except for some trailing stop that are based on curent market price.

I attach here the results of my test (on every tick).

How do you evaluate the potential of such a strategy? if you have examples do not hesitate!

I have seen you blog it looks interesting, I dont have time now, but I'll take closer look in the week ends!

Thanks again for helping me!

Best regards,

Patrick

 

DMP

This is marginal for use on a live account as the Profit Factor is a bit low, the Drawdown a bit high and the max consec loss is a bit of a concern

Also the number of trades (per year) is a little low to be statistically valuable

There are no absolute levels which are go/no go, you just look at the whole picture for areas of weakness

Most systems that are OK on backtest of live data or forward trading of a demo account experience a worsening of results on live trading

This can vary from 5% to 60% difference for the worse...

In general, a low PF would be less bad on a frequently trading EA, but the max consec loss would still be a concern...

Much depends on the type of signals you are using and results will also vary on different brokers

FWIW

-BB-

 
Demeuterpatrick:

Thanks Paul,

My orders are based on Close except for some trailing stop that are based on curent market price.

I attach here the results of my test (on every tick).

How do you evaluate the potential of such a strategy? if you have examples do not hesitate!

I have seen you blog it looks interesting, I dont have time now, but I'll take closer look in the week ends!

Thanks again for helping me!

Best regards,

Patrick

Thanks Patrick. BarrowBoy makes some good points. You haven't answered my first question though, which is the most important: are you are evaluating the EA only on the equity curve after optimisation? If so, then you have missed a critical opportunity to check for potential future profitability. You say that the curve starts at 2000, so one sound approach is to optimise from 1/1/2000 to 31/12/2008 (say), then run the EA on the optimised settings from 1/1/2009 to present ("out of sample"). Read: "if I had started trading the EA live in 1/1/2009 would it have been profitable in 2009?" Even this approach is nowhere near foolproof, but at least it gives some idea of whether you have curved-fitted.


BarrowBoy's point about maximum consecutive losses is my biggest concern: would you have the guts to trade serious money on a system that has been backtested to lose 18 times in a row, so you can expect to see somewhat more than that in live trading?


 

Wow, thanks for your input guys!

I see I need to optimise a bit more héhé.

What PF would you recommend? What do you achieve in your EAs?

Concerning the consec loss, what kind of code can you put in place to reduce it?

My EA is based on five EMA's + 3 BollingerB%.

Each order is divided in a "Target Profit" and a "Traling Stop" part.

Are you guys professional traders?

Bye!

Patrick

 

D

> My EA is based on five EMA's + 3 BollingerB%.

Ah ha - that explains alot, this strategy needs clear & sustained trends, hence winning better mid-07 to end 08 and mostly bad news since Dec 08...?

> What PF would you recommend? What do you achieve in your EAs?

1.5 is OK for a frequent trading EA, for the others I prefer 2.0-2.5 and 4.0 is achievable

> I see I need to optimise a bit more héhé.

Would not help - you would boost some periods, make worse on others

Try looking at setting TP & SL based on the current ATR or percentages thereof

> Concerning the consec loss, what kind of code can you put in place to reduce it?

Develop a better strategy - sorry!

Maybe add some filter to detect a ranging or at least non-trending market, maybe using ADX or VHF indicator

> Are you guys professional traders?

Yes

-BB-

 

Awsome!

Thanks a lot... ! I'll work on it

Really appreciate.

BTW, to get a lot of trades I guess we need to reduce the trading timeframe, Im now on H1, would a shorter timeframe be advisable? (of course the strat adapted to it).

Best,

Patrick

 
Demeuterpatrick:

BTW, to get a lot of trades I guess we need to reduce the trading timeframe, Im now on H1, would a shorter timeframe be advisable? (of course the strat adapted to it).

A shorter timeframe is the usual way to more signals - if testing shows the strategy or an adapted version is going to work

IMHO, I wouldnt go less than M15 if Bolli Bands are in use

FWIW

-BB-

 
BarrowBoy wrote >>

A shorter timeframe is the usual way to more signals - if testing shows the strategy or an adapted version is going to work

IMHO, I wouldnt go less than M15 if Bolli Bands are in use

FWIW

-BB-

Ok roger that!

Man, Just to have an idea, (no need to answer if the question is displaced), but How much do you make per month?

My job sucks and Im concidering to invest some time in forex since I realy like trading (Automated trading is still kinda new for me)

Well thanks again for ya help!

Ciao

Pat

 

How much do you make per month?

Enough to put my daughter through college!

And next year i'll be working full time in Forex

Worst thing you can do is jump too soon - if you keep at it, you might get to earning something regular in a couple of years

Thing is, you need a serious account size to make enough to take out and live on the profits...

Just cos the job is tough is no reason to jump ship, fully prepare and prove before committing yourself.

With EA's, there is no excuse not to prove their (live) earning power over a significant period before depending on them to pay the rent!

Manual trading is clearly different, but those skills are not given to all either...

-BB-

New comment
Reason: