forex software

Create and Test Forex Strategies

forex software

Skip to forum content

Forex Forum

Forex Software, Forex Strategies, Expert Advisors Generator

You are not logged in. Please login or register.


Forex Forum → Expert Advisor Studio → Portfolio risk issue - need max nr of trades option

Pages 1

You must login or register to post a reply

RSS topic feed

Posts: 3

1 (edited by Patrick2 2018-06-20 08:55:26)

Topic: Portfolio risk issue - need max nr of trades option

Morning Traders,

With "Walk Forward" again another big step to generate stable EAs - great job!

However I see a similar big risk like Hedge EAs.

Example:
We generate a portfolio of 100 or have read also 10000 EA inside.

And now it happens in one kind of moment in a year, that 10-50 EAs Buy / Sell at one time.

Can we please upgrade the exports with "MAX TRADES" a nr to enter maximum entry / open trades please.

Or can somebody help how to ad it manually into the script ?

This is at the moment from a risk reward view the only an one big "black swan" that EAS Portfolio has if you work
with 100 or more EAs in one file.

Why you need 100-1000 in one script?
Simple - try to search/generate for very tight setups. EAS will find strategies where enters only 5-20 times a years.
Group them in one big portfolio  - but in worst case as above - you may enter with 10 or more at one time.

Thank you for your reply
Patrick

Re: Portfolio risk issue - need max nr of trades option

Just a little warning why this idea is flawed from my own experience (as I had similar ideas when starting out with automated trading 10 years ago):

1) If you limit the amount of trades for the whole portfolio like that, you are effectively not trading the strategies in the way that they have been generated by EA Studio. You are randomly skipping trades this way and the strategies in the portfolio can not enter,  or enter at different times than what they really wanted to, because they could not enter at the bar they wanted because of the limited amount of trades that are allowed to be open at once. That effectively will make the strategies trade completely different from the backtest, hence creating a completely different system than what EA Studio showed you in the single strategy backtest. Especially with such a low amount of trades per year and per system, you are NOT wanting to miss an entry because of such a limit, because it can mean missing the few winners that such a low-frequency strategy creates each year.

2) Apart from that, strategies that trade just 5 to 20 times a year on their own (not combined in the portfolio) do not have any meaningful statistical edge. Such a low frequency of trades, especially if using just a few years of history, so that the total trades in such a strategy are only ~100 to 200 in the whole backtest, are almost guaranteed to fail. My experience shows that you´ll need at least ~1500, better 2000 trades in the whole backtest for a SINGLE strategy to really know that the results are not coming from random luck.

If you feel that your portfolio has too many trades, it would be better to limit the number of strategies in the portfolio, raise your capital, or use a cent-account broker instead.

3 (edited by hannahis 2018-06-21 22:20:24)

Re: Portfolio risk issue - need max nr of trades option

1) I agree with Geektrader that if you place an "external" instruction to limit your trade order/numbers, you have to be prepared that you are entering into "uncharted" territory, whereby you can't rely on your backtest to predict what the outcome will be like.  Nevertheless, that should not stop one from trying cos you may be surprised that such intervention may turned out to be a better option.  Likewise, I always add an external profit trailing EA to handle all the opened trades and hence, I can't refer to FSB backtest (since the backtest result can't take into account of my external EA interference, nevertheless, I get far better trade performance and hence, I'm willing to go "blind" (without any backtest results to refer to).  Assurance (having a backtest results to rely on) vs Performance (improving your trade results).  As for me, I'll do whatever external interference I could, if I know it will improve my trade results.  Cos I never let (the lack of) backtest dependence, limit the way I trade.  Ultimately, it's the trade results I'm looking for.  Whichever method that can lead to better results, I'll take that path.



2) We use different method to build our portfolio.  a) Portfolio with diverse EA/systems/strategies b) Portfolio with similar strategies c) We are clueless of how these strategies work, since we don't build it ourselves (via Generation, etc)

2a) Portfolio with diverse EA.  In order to put together a portfolio of diverse EA, I'm assuming that one has to know the trading behaviour of the EA in order to know which EA to select for the portfolio.  If we are well versed of how our strategies worked together, then adding external "intervention" such as max lot size allowed to be traded etc will deviate the Portfolio overall performance (from it's backttest "predicted" outcome).  In this case, if we do know how our strategies work, we would also have a better idea whether it is beneficial to have external trade limits or not.

2b) If portfolio are made up of very similar strategies and since we don't know which will trade 1st etc.  Then using a trade limit, max number of allowed positions or max number of lot size limit would be a good idea.  Eg, I'm prepared to risk certain percentage of my account and hence I limit the max number of lot size allowed to be trade and whichever (more responsive) EA open strategies 1st, I'll trade.  It's like 1st come, 1st serve.

I do sometimes use this method because I do want to control my level of risk factor and since I don't know whether all my EA will trade or only some of my EA will trade.  Using a max lot size being trade will give me a peace of mind.

I normally use FxBlue copy trade and hence, this copier has many features and one of the features allow me to limit the max number of lot size being trade or stop trading when account reached certain level etc

2c) For those who generate EA, we would most likely be clueless of how and when our strategies will trade, whether frequently or once in a blue moon.  Hence, max lot size limit may be a safe way to trade, bearing in mind that any external interference we use, will deviate from the backtest results.  Hence, it's a case of Peace of Mind vs relying on having a Backtest "assurance", if this is important to you.  I understand, some users are very systematic with their EA development and won't deviate anything from their backtest in order to ensure a workflow.  Hence, it all boils down to one's approach in EA development, there is no right or wrong.  It's a matter of personal preference, we do what works best for us.

In conclusion, it is important that we do know how our portfolio performance in terms of it's trading behaviour, what's the max number of opened positions overall, what's the max DD, etc before we ever go live.

As for me, I tried different methods 1) use max lot size allowed and 2) like what Geektrader suggested (1. Limit the number of strategies, calculate what's the potential max DD of the portfolio and adjust according to your risk appetite or 2. Increase your capital, thus increase your "allowable" Max DD, 3. Use percent lot sizing so that your trade are automatically adjusted according to your capital.

Posts: 3

Pages 1

You must login or register to post a reply

Forex Forum → Expert Advisor Studio → Portfolio risk issue - need max nr of trades option

Similar topics in this forum