Usdcad, audusd, usdjpy, nzdusd

Gentlemans. Does anybody found robust strategies on 30min tf with ducas copy data? I have with theese pairs problems from beginning. If yes could you provide few details related to builder settings and analysis filters? Thanks in advance.

I can confirm this, regardless of timeframe, SQ has a VERY hard time to find strategies for USDCAD, EURCAD, AUDUSD, EURAUD, USDJPY and NZDUSD, you are right with this. Hope SQ4 is a bit more advanced and will find some strategies on those pairs. What it shows, however, is that those pairs are the less stablest pairs in terms of repeating patterns which are needed for a automated trading system.

I can confirm this, regardless of timeframe, SQ has a VERY hard time to find strategies for USDCAD, EURCAD, AUDUSD, EURAUD, USDJPY and NZDUSD, you are right with this. Hope SQ4 is a bit more advanced and will find some strategies on those pairs. What it shows, however, is that those pairs are the less stablest pairs in terms of repeating patterns which are needed for a automated trading system.

I haven’t looked at M30 (I’ve looked at H1 and M5 mostly) and can’t find anything good on USDCAD or USDJPY and I’ve spent months on it.  AUDUSD I’ve found suspiciously few.  Pairs Like EURUSD and GBPUSD I find thousands that look good by contrast.

Not looked at the others you mention yet.  I suspect that these pairs are much more random in their behaviour as you say.

Hi,

Thanks for confirm my discoveries. In next week ill try build strategies on higher time frames on theese pairs. Will let you know.

I don’t think it’s the pairs, I think it’s SQ. EURUSD which it finds strategies pretty easily on is considered one of the noisiest pairs in the last 5 or so years because it’s heavily over traded. Several of the pairs mentioned have far less intraday noise. EURUSD just happen to have a good run the last year though.

It is the pairs, the mentioned pairs have very low repeating patterns compared to EURUSD, so it´s hard to find strategies that do well at all. You can check with data-mining tools, they find the same in relation to that the above pairs are a lot more unpredictable than EURUSD.

Interesting, 4-time World Cup trading championship winner Anrea Unger who is a 100% Quant (but writes human-made systems, not data-mined systems) says the complete opposite of you and I agree with him as far as non-data mined systems go, EURUSD has the most noise of any pair.
The more heavily a product is traded, the noisier it becomes, the less inefficiencies it has. S&P500 is a perfect example. That’s why smaller country’s stock markets trend far better and the US markets have far better mean-reverting potential. I believe that’s how Ray Dalio of Bridgewater actually allocates his systems.

Not to come off topic though, I find SQ is much more successful with the EURUSD. I hope SQ4 has potential for smarter custom searches to take advantage of other pair’s personalities.

Look, might he have won 100 world-cups, I meanwhile create and trade what makes me the money… and that clearly is EURUSD as it comes up with the most stable systems, fewest movement changes. It might be noisy on the lower TF´s, but the larger timeframes tend to repeat history a lot more than any other pair I have traded in the past 10 years (that´s also my experience when I coded sytems without SQ before) since I am doing this full time. And honestly, I couldn´t care less about what anyone else says as long as this thing keeps on spitting out good EURUSD systems that constantly make me money going forward. Don´t know about you and if your goal is to come up with theories that might be true or not, or if you want to make money… I am definietely here to make money and share my finds on the way, that´s all I care about.

The reason I say this is because the systems I make that are not data mined tend to have the best success on the ones listed plus most of the crosses and they are all D1 whereas on EURUSD I’ve had to tweak or add parameters to make them work usually.
I think the systems SQ produces (which all seem fairly similar- most using very simple entry criteria with buy/sell stops put X-distance away from market based on volatility in essence “breakouts”) seems favorable to EURUSD’s personality but the other pairs definitely have patterns and personalities to them that can be taken advantage of, but might require more of a human eye, or in our case a bit more generating.

I disagree and some very intelligent and rich billionaire hedge fund managers have stated exactly the same about US markets that I have but I dont wish to debate it to no end. Its pretty obvious that the more members a market has, the more patterns are exploited and filled, and the more often it will reach sideways equilibriums (larger sample size of people more likely to reach split decision/disagreement more often than a much smaller sample). That’s why penny stocks and Bitcoin are so explosively volatile. Pairs like EURAUD and GBPCAD have some of the best trends but just recent fundamentals have gave the dollar a huge push while the euro falls apart.

I’ll simply leave it at that.

What’s  interesting to me, is you are both obviously intelligent people with experience of the markets, but you are at odds about EURUSD vs less traded pairs (crosses and minors).

On one hand EURUSD is the most liquid and heavily traded pair, with the best spreads I would suggest.  One the other hand, such heavy trading must influence the structure of the pair, possibly to the detriment of market inefficiencies.

I don’t know what the answer is, but using SQ I definately find what looks like very good systems.  Same with GBPUSD.

Happy pips to all.

1 By sideway equilibrium I suspect you mean congestion since supply rarely equals demand.  I try to be specific rather than use very difficult to prove generalisations such as “some very intelligent and rich billionaire hedge fund managers”. I think you’ll find George Soros has a contrary views to your position…check out his theory of reflexivity which he attributes to his outstanding trading success;

 

2 Fibonnacci tools are one of the most widespread tools used by professional and retail traders; I assume most traders  (from noob to pro) know the basic tenets of these tools and that they work due to crowd behaviour and the larger the crowd the better they work.  See: Khoshnevisan, Mohd, and Soukanto Bhattacharya. “Financial Analysis and Trading: Financial Engineering and Applications of Fibonacci Sequences.”ADVANCES IN GLOBAL BUSINESS RESEARCH (2012).

 

3 “larger sample size of people more likely to reach split decision/disagreement more often than a much smaller sample” That statement is simply asinine.  If you have a sample of 1000 people randomly saying yes or no then the EXPECTED VALUE of a yes or no being said is 50%.  The expected value is 50% regardless of the sample size. 

I didn’t realize this was such a sensitive topic to produce inflammatory remarks like “asinine”.

1 Ok, So I’ll have to ignore your quotes then or I’d have to fill the page with quotes from the likes of Jerry Parker, Ed Thorp, Ray Dalio.

2 Notch do you have any mechanical systems running fibonnacci proving its statistically relevant and predictive? Can you predict if its going to retrace to a 38.2, a 50, or a 61.8 or prove that it was even the fib that caused the bounce or it was it Support and resistance from previously defined levels or overbought/sold indicators, or actually *a confluence of EVERYTHING*? I use it discretionarily on a daily basis because I combine it with so much more. This is not a consistently and mechanically repeating pattern or tool that can be used on its own or even be said to be a sole contributor to the patterns it appears to be a part of therefore your point is irrelevant. This is actually a great example, I’m so glad you brought it up, because the majority of traders DO look at it and the smart money knows this. That’s why its levels are constantly whipsawed, broken out of falsely, or just plainly ignored. Breakouts of SnR, donchian, whatever for example, like the turtle system were absolutely killer in the markets until they got over used and the smarter money took advantage of it to cause more frequent false breakouts in today’s market (especially forex where more volatility filters are required to find the right set ups for true breakouts).
You’re either trolling with much of this acting deliberately inflammatory and playing ignorant, or truly happen to be this way.

3. Really? So you feel fine with 2 robustness tests because they should be statistically relevant? Comfortable with 2 coin flips versus a 1000? Don’t think it stabilizes the statistics? Ok then…. What you’ve just said is a complete fallacy full of ignorance (I feel its from you trying to be a troll and actually slipping up) and is indeed the asinine remark. Smaller samples *destabilize* the statistical balance/relevance. I don’t want to misinform any new traders. Ex: If 2 Traders will more often take the same side than 200 traders. That’s like statistics class 101 dude.

Tulip Mania and lets add bitcoin. These were/are young markets, that had yet (in BTC’s case) to become heavily systematized. Markets becoming more “erratic”/”random” than ever before is a phenomenon produced by mechanization. Forex is considered the “smartest” market in the world because it is the LEAST emotional and it is highly systematic. Many patterns have been exploited in it because it is saturated with a variety of systems (most symmetrical) which give more balance to supply/demand (thats why we’re here randomly generating to find them, aren’t we? 😉 ). Do I believe its efficient? No, but EURUSD is the most of any pair followed by USDJPY, and I stand by what I said. Just because a data mining software finds strategies in 1 pair vs another doesn’t mean the other pairs have less statistical patterns, it entirely depends on the structure of those strategies it produces paired with the asset that’s being tested, more so than you’re giving weight to.

My position adds even more reason than yours to why we’re all here using random generation and Marc’s amazing software suite to find patterns in the market- Because finding patterns in the market (especially the most efficient one:FX) is HARD! Thank you. 😀 Notch, I enjoy your posts and learn from everybody on the forum, but I enjoy the peace on the forums more. All morning you’ve added intensity with certain remarks and tone that was not here before.

I haven’t looked at M30 (I’ve looked at H1 and M5 mostly) and can’t find anything good on USDCAD or USDJPY and I’ve spent months on it.  AUDUSD I’ve found suspiciously few.  Pairs Like EURUSD and GBPUSD I find thousands that look good by contrast.

 

Not looked at the others you mention yet.  I suspect that these pairs are much more random in their behaviour as you say.

MikeyC i found some good streategies on USDJPY 1H TF. Here is my settings

Guys, relax, why not just spend the time on generating systems instead to waste it with what should work in theory and what not? In the end the one that makes more money with her/his approach of generating systems in SQ is the one who is right, nothing else really counts nor does matter.

After many month of trial and error i stoped searching for new strategies , I beleive that the min requirement for a system ( strategy) to work out is at least analysing 2 timeframes at the same time . 

it is impossible to scalp or swing on a low timeframe 1hr and below without checking the all over trend in another higher timeframe like day or h4 at least  ,

Save your breath and wait for sq4 , especially that the multi core will be fixed . 

The first important factor of GA is speed then comes parameters or indicators , speed was the first issue from when GA was created , 

Can you not “see” the longer term trend on lower timeframes by using the appropriate (large) period settings on some indicator?

How do you currently decide, for example the overall trend on daily chart?