Topic: live trading not looking good

Hi good traders
So after watching all Popov's video tutorials on how to generate and test trading strategies with the Forex Strategy builder professional , I generated about 12 different strategies for EURUSD and another 12 different strategies for USDJPY on the 15minutes chart timeframe.Why so many different strategies?Well Popov suggested in his videos to create multiple strategies(portfolio) so that they can compliment each other(diversification) ..After generating the strategies I tested them in Monte Carlo simulation tool and I only cherry picked the strategies that passed the all the monte carlo simulations as seen in the screenshot.And I only picked strategies with 95% confidence level of doubling the account....

After monte carlo simulations , i decided to run a test with the portfolio of EAs in a small live account.Same broker historical data i used to generate the strategies...After 2 weeks of live test the portfolio has made a loss and is in drawdown about 20% currently.Meanwhile the drawdown of porfolio in the forex strategy generator software is 31%  and the maximum stagnation period is 6 days...Its been 2 weeks now since I started testing and the portfolios are it has made a loss and in drawdown .Doesnt look like the maximum stagnation is 6 days as the forex strategy builder professional portfolio says in the portfolio I generated with the software.

I dont know what I did wrong and dont know what to do because the statistics in the portfolio are diferent from the statistics in the live test.
I will be glad if someone can point me in the right direction.Thank you for reading this.

Post's attachments

MonteCarlo.png 87.05 kb, file has never been downloaded. 

You don't have the permssions to download the attachments of this post.

Re: live trading not looking good

I am not sure what your portfolio looks like so I can not comment on it.

My personal approach is to get a number of strategies that have 75 or 85 win/loss and weed out the poor performers on a regular basis.. using the MT4 Tracker that Sleytus developed.

I keep replacing the poor performers.

I guess I have over simplified things.

Regarding the tapes and instructions...... those are to help people to learn to use the software... they are not a definitive guide to automated trading. We have to learn that from our own education and experience. We have to try things until we do everything we can think of to make a strategy work.

I suggest that you keep your strategies very simple and ensure that they have high metrics, be quick to ditch a poor performer.

The market changes its face at random, there is not patterns that we can always depend on, we have to be nimble.

I get around the stagnation by having a lot of different strategies, so they are opening trades at different times.

I use the smallest lot size my broker offers....., increasing SLOWLY as the account builds.

Keep in mind that patterns that worked  two years ago may not be viable today.

My last idea is......... did you do walk forward on your data?

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: live trading not looking good

Usually when I am reviewing a strategy, I look at the thing on the chart in FSBPro to see if I can see myself trading it manually...

I try to get a feel as to how I would react as the signal appears.

If it is too weird, I pass on it.

I also try to avoid much in the way of drawdown, I use tight stops. I am afraid of several strategies hitting the stop loss at the same time so I try to keep the stops small.

I trade manually each week for quite a few hours, it gives me ideas for ea's and also shows me weaknesses in my thinking. Sometimes we think that automated trading is the only solution and stay away from manual trading. I like to keep my hand in the manual trading, so that I stay in touch with the market.

The Monte Carlo chart looks great. I can not tell much from that as I do not know what the criteria is for your entries and exits.

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: live trading not looking good

Thanks for the recommendations.I do not use much indicators when am generating strategies.I use maximum 4 indicators.I set the FSB software generator settings this way "Maximum opening logic slots=2 and minimum opening logic slots=2 "
.I have not done walk forward analysis.How do I do that in the SPAM? When you say you keep replacing the poor performers how exactly do you do that?Do you remove the poor performers every week , everyday , every month or what?How exactly do you do that?Thanks for the tips and recommendations.very much appreciated.Bless you and green pips to you Good Sir.

Blaiserboy wrote:

I am not sure what your portfolio looks like so I can not comment on it.

My personal approach is to get a number of strategies that have 75 or 85 win/loss and weed out the poor performers on a regular basis.. using the MT4 Tracker that Sleytus developed.

I keep replacing the poor performers.

I guess I have over simplified things.

Regarding the tapes and instructions...... those are to help people to learn to use the software... they are not a definitive guide to automated trading. We have to learn that from our own education and experience. We have to try things until we do everything we can think of to make a strategy work.

I suggest that you keep your strategies very simple and ensure that they have high metrics, be quick to ditch a poor performer.

The market changes its face at random, there is not patterns that we can always depend on, we have to be nimble.

I get around the stagnation by having a lot of different strategies, so they are opening trades at different times.

I use the smallest lot size my broker offers....., increasing SLOWLY as the account builds.

Keep in mind that patterns that worked  two years ago may not be viable today.

My last idea is......... did you do walk forward on your data?

Re: live trading not looking good

As far as the Monte Carlo simulation is concerned I always follow the exact rules by Popov when running my Monte Carlo simulations..i wonder if you or anyone else is making consistent profit in live/real account from the EAs generated from the FSB?

Blaiserboy wrote:

Usually when I am reviewing a strategy, I look at the thing on the chart in FSBPro to see if I can see myself trading it manually...

I try to get a feel as to how I would react as the signal appears.

If it is too weird, I pass on it.

I also try to avoid much in the way of drawdown, I use tight stops. I am afraid of several strategies hitting the stop loss at the same time so I try to keep the stops small.

I trade manually each week for quite a few hours, it gives me ideas for ea's and also shows me weaknesses in my thinking. Sometimes we think that automated trading is the only solution and stay away from manual trading. I like to keep my hand in the manual trading, so that I stay in touch with the market.

The Monte Carlo chart looks great. I can not tell much from that as I do not know what the criteria is for your entries and exits.

Re: live trading not looking good

How long have you been doing this....the reason I ask is not to point out your experience in FX market but to highlight that when you create a strategy you do notice that there are drawdown and stagnation periods....looking simply at the graph..we see it doubled the account in a year and we put it in a demo/live account...

If there is a drawdown of 30% and there is stagnation for 20 days...then it is very possible you entered the market at that time...

Lets say that in your backtest you started with $1,000 in Jan 2017 and by November it shows account doubled....however in April there was drawdown of 30 percent. Now imagine you had entered the market in April with this EA...in the end it will make money but a person standing in April would be looking at their account with a loss of $300.

This happens...I have seen people do a backtest on 10 years and not realizing that in the last 6 months the EA did nothing for them (in the backtest)..the graph looks amazing though.

I hope you can make sense of my rambling...

Re: live trading not looking good

araza wrote:

I hope you can make sense of my rambling...

I think you explained it perfectly.  And that is also my experience.

8 (edited by Irmantas 2017-12-12 11:33:00)

Re: live trading not looking good

Hi,
You will need to find out right path to success by yourself. There is no single right path and there is ton of losing ones. Best thing in my opinion is to start testing different ways of making things (of course with demo money, or using OOS) and see what works for you. Also I recommend to make some study if you are green in this area. There are some great books etc. :  https://www.amazon.com/Trading-Systems- … ni+trading, https://www.amazon.com/Building-Winning … GKTBSFYFWR, http://www.priceactionlab.com/Blog/the-book/ . Also you can listen some interesting interviews: https://www.youtube.com/channel/UCdnzT5 … TOiDsPhqcg, https://www.youtube.com/channel/UCXtFh3 … ORL1tZxKrQ.
By studying this stuff you will find variety of different ways how people make automated trading work.

Try to answer these random questions:
How much data/bars is right?
How much indicators/variables is OK?
How to optimize strategies?
How much trades is OK?
How much return/risk expect from strategies?
What is turning off point of single strategy or portfolio?
What is data mining bias / curve fitting?
Is generated or manual created strategies better?
Does out of sample testing makes things better?
What is walk forward analysis and does it make better systems?
What is System Parameter Permutation ?
What shortfall from a backtested results I can expect when I go live?
... and in the process you will find this list endless with more questions what needs to be answered. And probably in the way you will find out what really matters and works smile Also by asking some questions you are forced to make experiments, and some great finding awaits you.

For me, automated generated strategies on short time period did not worked out like you. But I found some successes using more data, and using OOS check. Also I got some success with some manually created strategies.

Re: live trading not looking good

thanks for your contribution..Appreciated..

araza wrote:

How long have you been doing this....the reason I ask is not to point out your experience in FX market but to highlight that when you create a strategy you do notice that there are drawdown and stagnation periods....looking simply at the graph..we see it doubled the account in a year and we put it in a demo/live account...

If there is a drawdown of 30% and there is stagnation for 20 days...then it is very possible you entered the market at that time...

Lets say that in your backtest you started with $1,000 in Jan 2017 and by November it shows account doubled....however in April there was drawdown of 30 percent. Now imagine you had entered the market in April with this EA...in the end it will make money but a person standing in April would be looking at their account with a loss of $300.

This happens...I have seen people do a backtest on 10 years and not realizing that in the last 6 months the EA did nothing for them (in the backtest)..the graph looks amazing though.

I hope you can make sense of my rambling...

Re: live trading not looking good

thanks for the tips.Very much appreciated..

Irmantas wrote:

Hi,
You will need to find out right path to success by yourself. There is no single right path and there is ton of losing ones. Best thing in my opinion is to start testing different ways of making things (of course with demo money, or using OOS) and see what works for you. Also I recommend to make some study if you are green in this area. There are some great books etc. :  https://www.amazon.com/Trading-Systems- … ni+trading, https://www.amazon.com/Building-Winning … GKTBSFYFWR, http://www.priceactionlab.com/Blog/the-book/ . Also you can listen some interesting interviews: https://www.youtube.com/channel/UCdnzT5 … TOiDsPhqcg, https://www.youtube.com/channel/UCXtFh3 … ORL1tZxKrQ.
By studying this stuff you will find variety of different ways how people make automated trading work.

Try to answer these random questions:
How much data/bars is right?
How much indicators/variables is OK?
How to optimize strategies?
How much trades is OK?
How much return/risk expect from strategies?
What is turning off point of single strategy or portfolio?
What is data mining bias / curve fitting?
Is generated or manual created strategies better?
Does out of sample testing makes things better?
What is walk forward analysis and does it make better systems?
What is System Parameter Permutation ?
What shortfall from a backtested results I can expect when I go live?
... and in the process you will find this list endless with more questions what needs to be answered. And probably in the way you will find out what really matters and works smile Also by asking some questions you are forced to make experiments, and some great finding awaits you.

For me, automated generated strategies on short time period did not worked out like you. But I found some successes using more data, and using OOS check. Also I got some success with some manually created strategies.

Re: live trading not looking good

When I get a 'wonderful' strategy, I have a look at the journal..... observe the account balance column... looking for drops....., look at the trade dates....... looking for gaps. Have a look at the 'Balance Chart'. I have quite a few different strategies, I am not too concerned with stagnation. ( the tighter the stops and take profit, the smaller the problem with stagnation)

Then I use the 'walk forward', it is really simple in FSBPro, just change the dates and see what happens in each period.

Keep things really simple and observe the different tools to detect problems.

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....