Topic: Robust Trading Strategies

We often have questions as to the viability of the strategies we are creating, whether or not we can trust them to make us a profit.

I have gathered a few quotes and will share them with you.

Before I do that...... I want to point out that the "Multi Markets" Tool on the Analysis Page of FSBPro is the 'whole truth' re robustness of a strategy.

Using "MultiMarkets" is fast and efficient for determining the robustness of your strategy.

There is not much need for any other tests as you can test various instruments and all time frames simultaneously using this tool. It takes a few seconds to compute. and the results are vividly displayed.

Of course, simple strategies using one or two logical slots are the ones that will make the grade. If you load up the logic slots with 5 or 10 indicators, chances are zilch that your strategy will survive.

******************

Stable trading systems tend to be longer term in nature.

A robust system is one which works and is stable over many types of market conditions and over many timeframes. It works in German Bund futures and it works in Wheat. It works when tested over 1950-1960 or over 1990-2000. Robust systems tend to be designed around successful trading tactics (origin of our “Tactical” name), classical money management techniques, and universal principles of market behavior. These systems are not designed around specific types of markets or market action.
Dr David Druz

To design any mechanical strategy it is important to consider three things before anything else: 1) your objective for that system, 2) your market, 3) your timeframe. Once you have determined this, it is easy to find your essential methodology because there are only 4 ways to trade any market: 1) trend trading, 2) momentum trading, 3) reversion to the mean trading, 4) and fundamental trading. Once you have determined your objective, market, timeframe and method you are ready to attempt to put together your first strategy.

Brett Steenberger

build a strategy over multiple markets, where each market is different. Some traders only trade multi-market strategies based on the belief that single-market strategies are too likely to be over-fit.
Michael Bryant

In summary a robust system should do the following

1. Trade a large portfolio of markets successfully
2. Trade that large portfolio successfully over a very long test period
3. Use the exact same rules for every market
4. Use the exact same input values for every market even if the rules are the same
5. Have the exact same logic and input values for initiating both buys and sells
6. Factors in realistic transaction costs
(slippage & commissions)
7. Be tested in a way where the markets have been normalized for risk
(not single contract)
8. Doesn't use Static preset exits for all markets IE $2000 stop or $5000 profit target for all markets, but rather dynamically computed ones.
Dean Hoffman


Practically speaking, a robust trading strategy is one that produces consistently good results across a broad set of parameter (input) values applied to many different markets tested for many years.”   Perry J. Kaufman

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

Re: Robust Trading Strategies

Practically speaking, a robust trading strategy is one that produces consistently good results across a broad set of parameter (input) values applied to many different markets tested for many years.”   Perry J. Kaufman

Hello Blaiserboy

basically you're saying that an EA (with its strategy), operating correctly, • must work even if I change its input parameters
• on different cross
• 50+ years

I think it is the search of the impossible, have you already found a similar strategy?
It is better to fit a base in the market strategy, which look something iraggiungibile? at least with these tools.

Re: Robust Trading Strategies

Well, as you gain experience with FSBPro, you will gradually get to develop strategies that can work on several markets.

Can I suggest that people like Kaufman have the experience to make such statements and qualify their opinions.

I have posted one strategy that I use... and since posting that I have made several similar.

My strategies are based on trend and breakouts, they do not get a lot of trades on a daily basis, they are all higher time frames.

My data starts in 2001.

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

Re: Robust Trading Strategies

If i may ask, what are the factors you look at mainly when creating your EA

e.g max draw-down, profit factor etc

Thanks

Re: Robust Trading Strategies

Hey

I like to see SQN above 2.5..... to start with

Profit factor above 2

Win Loss above 60

Drawdown no more than 60

Those are my preferences...... each of us has to establish his personal settings... some can take more risk than others.......

Hope that helps

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

6 (edited by Don-Essa 2016-07-23 18:11:29)

Re: Robust Trading Strategies

Blaiserboy wrote:

Hey

I like to see SQN above 2.5..... to start with

Profit factor above 2

Win Loss above 60

Drawdown no more than 60

Those are my preferences...... each of us has to establish his personal settings... some can take more risk than others.......

Hope that helps

Hey Blaiser boy

Thanks for the feedback it is very helpul as i can also try and determine my parameters as well. when you mentioned win loss above 60 did you mean 0.60 or 60.00?

How about the initial amount that is set on FSB Pro.......do you like usually put the initial amount based on the same amount would be in the account that you would be trading in weather on demo or Live account.

E.g: I use $10,000 as the initial amount set under the control panel in FSB pro and then use it on an account that reads $100, would this be something to consider strictly?

Thanks

Re: Robust Trading Strategies

I use $300 for my testing, sometimes I use $100...  I do not know what is right to do.

I spend most of my time trying to develop strategies..... and those amounts seem to be ok for me.

One proviso is... if your entries are percentage of account based. I suggest that you use an amount close to your account as the outcomes can be very important as lot sizes fluctuate.

Another proviso.... use only a small portion of your funds in your real account. you have to have lots in reserve for a rainy day.

I know I did not give a great reply to your question... you will have to research and develop your own idea.

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

Re: Robust Trading Strategies

I think this is a very good topic!


First of all, I want to give a feedback regarding FSB, this software is simply responsible for tons of profit...
I started using the software May 2016.

Basically the points you mentioned on the first post are the search for the impossible for any timeframe less than 4H, after zillions of hours generating creating and optimizing EA's its 100% impossible to have this DNA with 15M, 5M or 1M timeframe.
Then a critic point comes to my mind, when do I leave behind an EA and put it to a trunk/trash? How many new EA's we create based on last one or simply create 100% brand-new, consequently we start creating dozens of new EA's.
I started trading institutionally equities in Brazil in 2012 and all of this FOREX markets are new for me, we are used to manage and optimize portfolio of stocks under the same trading logic, trades that lasts few days others starting on the market opening and leaving it at the end of the day.

Let me bring this community few questions:

1) How healthy is to keep creating new strategies instead of bringing more complexity or more filters to the ones you already have? Why everyone seems to create 129381239 instead of only 2-3 that are their "biggest champions"?

2) I noticed on commercial EA's that they optimize it from times to time, how often do you recommend optimize an 1M,5M,15M?

Re: Robust Trading Strategies

I will give you some ideas that I have, they are not the gospel, just my approach.

Over a few years I have developed a number of strategies for daily and 240 minute charts, only a couple indicators in each. I optimize them once in a while, maybe every 6 months or so. I use data back to the mid 90's, so 6 months does not add much data.

The reason I use the higher time frames is that they do not need a lot of attention, once in a while I will weed out a poor performer.

I used to be very picky about my strategies, using strict parameters. now I am not so sure I did the right thing. Now my approach is less strict and I have many more strategies trading.

At this time I am generating 15 minute strategies...... many of them, I watch them and weed out the poor ones as weeks pass... I add quite a few, ie 30 or 40 each day, I have a few hundred running. I am not really concerned about optimizing them, some will make the grade and some will fail. The big determinant seems to be the indicators used, some indicators are too sensitive causing the strategy to fail. Too many indicators can be huge trouble. I am not convinced yet that a 15 minute chart is worthwhile for my purpose, I may shift to 60 minute soon..

Regarding 5 minute and one minute.... I have not worked with those at all, I would think they have to be optimized every thousand bars.

There is a couple thoughts.... hopefully others will add to this.

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

Re: Robust Trading Strategies

How do you come up with names for all those EAs Blaiserboy?!

Re: Robust Trading Strategies

Blaiser,

We're basically on same points, you are so right about the lower timeframes charts. But I don't know if its because of my equity-portfolio-handling background it doesn't sound right to simply throw away EA's because they started to "fail", what is considered a failure?

On quantitative approach I can consider failure as:

When a model compared to its past performance (OOS) simply wont correspond according to the trader goal(profit), means that this pattern has been lost for an unknown period, maybe forever.

But the tricky thing is: Lets say I backtested a strategy 15M timeframe and put it to run / simulate today (21/10/2016) and during 10 or 20 days it goes bad, how can I classify that model as failure? I dont have enough sample to do it and I simply discarded it, and that 10-20 days are just an exception and the rule is something TREMENDOUSLY profitable... The fact is that if you have 3-5 EA's running on test you can follow that but if you have 90-100 you just can't and oyu start to randomize your trades without sampling, and this results are all based on past statistics but then you start to ignore all of them.

Basically we have to be very careful when creating new stuff...because we must watch it trading check the trades on chart, it is very hard to do this checks when trading 30 EA's at same time

Re: Robust Trading Strategies

Well, I agree with you on all.....

I like to gather the experience of a walk forward for many trades to get a sense of what can be expected....., make a decision based on what I have done with FSBPro and also a walk forward period.  This is different from what I used to do.

As for discarding one that has gone bad for a few trades...... we have to check back as to what happened when we were testing it...... are the conditions causing it to fail today much different from what it experienced in our testing period.

Certainly I do not advocate throwing out the work done in the past, too many hours go into each strategy.

I wonder if you have a system of measuring performance as the strategies trade...... ie do you analyze the trading results on MyFXbook to see whatever that might show.

I use a couple tools to evaluate results from time to time......, sometimes something as simple as not trading on Monday can heal some problems with a strategy, especially with the lower time frames......

I do not have any idea as to the constraints you use, I like to limit day of week, time of day, volume, etc on lower time frames...., nbars exit on the higher time frames..

I do not think there is  a 'right' way to build the strategies and manage them..... it is an ongoing process and we will probably never get it perfect.

Perhaps we can ask Popov to create some sort of analysis program for trading results....... I know that he is swamped today, perhaps as time passes he will be able to help us with this aspect.

I would like to see something that is a bit intuitive to take the results and make plain the flaws and the potential for correcting those before we trash the strategy..

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

Re: Robust Trading Strategies

My setup is pretty much the same I had when trading equities:

3 environments:

Live environment

Simulation environment

Development environment

Live and simulation are VPS close to my datacenter, I use IronFX for very specific financial reasons and they are running EA's I am evaluating and Development is a XEON to optimize/generate new strategies quickly to evaluate my performance I use the signals page from MQL5 community that also helps me to follow my positions when I am AFK.

So basically I have dedicated metatrader to follow that simulation trades.
For the live ones I have the signal and I have a robust pc for the maths, I dont use FSB For trading, this way I export the EA's files and transfer them to my VPS machines, do you use FSB for trading whats the good advantage of it?
I have to admit that you are right about that part of analyzing, MT4 is very poor on that sense when we talk about 10 EA's to analyze.

Why the commercial EA's are so STABLE like forever? What kind of filters they use that we miss on our valued FSB, how do you start creating one EA? Using vanilla indicators or that super zing zang zeng chinese candle formatter?

Re: Robust Trading Strategies

Simon wrote:

How do you come up with names for all those EAs Blaiserboy?!

I use a numerical system

date developed
currency pair
Time Frame
etc

I think my number is 15 digits and the last 4 are the magic number

I find that easy although sometimes I make mistakes entering the magic number when attaching to charts. LOL

I usually send 30 to MT4 as a batch and attach them to charts while waiting for the Multi Tester to finish.

I think each person has to develop their own concept as to how to develop and manage their portfolio, depending on what they want to accomplish and how they can organize their approach.

I used to attempt to use the indicators and time frame but now I am using more than one or two time frames and indicators in different combinations so I developed that numerical approach.

Probably the thing will be unsatisfactory as time passes and I will have to use a larger number to include more things.

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

Re: Robust Trading Strategies

That sounds much more sensible than the way I have been doing it... which has mostly been like the method you would use to name a pet.

I will take a leaf from your book!

Re: Robust Trading Strategies

I saw some other naming schemes in another thread, but I do not recall the thread name. Some people use a spreadsheet to keep the data on each strategy.

It is something that you will have to think on until you devise something you feel good about

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