Diversified Trend Trading Approach

Thanks C. There is a huge flaw in fixed $ approach. I will explain in a little bit if you don't mind me raising some flags... it is also associated and relevant to the position sizing flaw in conventional schemes. Again, my focus is purely consistency because there is tremendous pitfalls to strategies based on pips or $. The outcome will always vary which we much try to avoid at all costs. I just want to get my thoughts and philosophies across and let you all decide on however you want to use it. I think it will help in portfolio modelling... more to come after I get some "actual work" done and wrap up my week.

I can only say:

Awwwwww dat's nice R, What a cute little puppie :-)

I can only say this....Thanks for Letting me Devour Your Souls :-)

Attached Image (click to enlarge)

As spooky as the image above is.....that is very much how this forum works in a positive way, and how we can build our knowledge base if we choose to.

Many are the posts in this forum that criticize the quality of content, the general noise, the multiplicity of bantering egos etc. etc. etc......but rarely do you hear about the positive aspects that can make this forum very addictive, inspiring and actually beneficial for your trading journey. I would like to impart some of this positive aspect of forum life to you guys.....as it really is powerful but is based on the mindset you take in approaching forum content.

I am in a very good place now in my trading confidence and general attitude to life which is accelerating me towards my goals....and so much can be directly attributed to you....my fellow folk of this forum....whether you accept it or not.

The interactions I have had with great guys (guys and gulls) in this forum....and the wealth of ideas spewing forth amongst the absolute crap is quite literally unbelievable and in my mind at least accelerates as opposed to decelerates your process. Not only have I gained lifelong friends in this space from those who choose to interact with me......but I have also gained lifelong power from those that don't....whether you like it or not....simply from reading the intriguing ideas and then testing them myself.

Yes....there are other more serious formal venues such as Linked In....but the real power comes from FF which is more informal in nature....possesses tons more crap....but also holds the hidden gems. Just like every aspect of life...you have to dig a bit to find them....and you can't be spoon fed this stuff. You need to first find it...recognise it's value which can be tough if you are sceptical.....absorb it and then test it.

That is how knowledge and experience works guys. That neuroplasticity of your brain is what really what this is all about. You are an amalgam of your experiences which can be traced directly back to how you choose to participate on this forum. You really are Borg and you are assimilated.

We are all buddies here.....there really should be no egos. That's why you quite literally won't find followers on this thread as this two way process of sharing and absorbing is how the door swings on this forum. If you frame your experience on this forum negatively....you will waste so much time here......and really you have to question why you choose to continue to me a forum member.........but if you flip that mindset to a positive and receiving one....as opposed to a taking one.......then light that rocket fuel and prepare to take off on an exhilarating journey.

The Yin Yang of this secret equation is that "The more you share, the more you learn". If we break this little nut down what it means is this. You will actually learn very little from the content I have posted on this thread from simply reading it as that is not how learning actually works. The content is sound however...so how can that be? The secret lies in the fact that learning can only be instilled through 'acceptance'. For that to occur you need conviction....and for that to occur....you need to test the ideas yourself. That is the final crucible to which all that content needs to be passed to have any chance of sticking in the grey mush we call a brain......otherwise the info is simply a 'wall of text' that has as much value as wallpaper......so now back to that little gem. What I have shared in my posts is my life experience in trading. It is what I have learnt by this process of amalgam of content. It demonstrates how my brain is attacking the issues of trading and the process of sharing is attempting to coherently synthesize my thoughts.

I am the net winner in this......but you can be also if you 'flip the mind' and start reading the content for how it is intended. It is simply text to the reader....but ideas for the more creative 'doer'. So go ahead....test out some of these ideas and watch your learning accelerate.

So cheers guys.....all of you. I have thoroughly enjoyed absorbing you all......and hopefully will continue to do so while this experience remains positive for me. :-)

C (your FF Borg)



PS ....and I would like to thank FF from my heart for creating this venue of 'power' :-)

Why Strategy Design by a process of applied logic beats optimization every time

You may be aware that at the moment we are going through the very labour intensive process of searching for new strategies to complement the portfolio blend. Many may be getting frustrated that this process takes lots of time when the power of optimisation can crunch through all this in the blink of an eye through a clever genetic algorithm. So why the big deal in following the turtle when you can jump on the back of a hare?



Well the long and short of it is that any strategy by its nature contain a number of variables that influence each other in different ways and degrees to produce a non-linear outcome. The more variables to a strategy, the more possible permutations of outcomes until very quickly you are left with a morass of outcomes that do not lead you any further in your query regarding the necessary part of 'understanding your system'. Now the crucial point to note is that each strategy and it's design constraints means that by and large there is no single guaranteed assumption you can make in terms of what works and what doesn't. The non-linear relationship arrangements between variables means that each design has it's unique weaknesses and strengths. THIS IS DEPENDENT ON THE RELATIONSHIPS BETWEEN VARIABLES MORE THAN THE VARIABLES THEMSELVES. The adoption of mantras in this process is your worst enemy. TEST EVERY ASSUMPTION and understand it's impact on the relationship between other variables.

You see this principle at work in optimisations using MT4. A small number of variables leads to a small number of various outcomes. With more variables, the outcomes expand to a huge number in a non-linear way. This is the result of multi-variable relationships as opposed to a simple 1 to 1 arrangement. You get a 1 to many relationship as your variables in your strategy increase. Very quickly you are left with a minefield of choices and the one you select is invariably curve fit.

If you understand the variables and their impacts and 'benchmark them' with respect to their singular impact plus their colluding impact with multi-variables then you have a hope of understanding where it all went wrong or all went right under live trading conditions.

Not only does optimisation lead to 'curve fit solutions'....but importantly it leads you nowhere in understanding the factors that shape the outcomes of your live trading experience.

The process of building strategies by progressive logic building addresses all that.....and while it takes a long time to achieve an outcome....it is always heading towards a more beneficial outcome. This logical progression towards developing successful strategies is what separates the chaff from the wheat in terms of your backtesting experience.

Without this logic driven process.....you are once more adrift on the seas of improbability without a rudder.

Cheers

C

Sharing end of week outcome from my post earlier just to highlight the fact that closing all positions on Friday as a policy towards minimizing weekend gap exposure is not always a rosy picture. That is what markets do. "Every action has a every and opposite binary reaction".
Either we less profits run and cut losses short OR we take frequent profits and then cut losses. Both are poisonous but okay if they are aligned with your goals and objectives to risk tolerance. As highlighted this month, taking a hit on equity on every Friday and give some decent profits acquired during the week makes me sleep well when markets are closed and I am never ever going to feel helpless at mercy of waiting for Sunday open to see if I got a margin call or not.

Question C. Is this the right professional approach in your experience? What do you do specifically to cover weekend risk (if any). What do the professional FMs do? I will share this though. Across Divergence and Convergence, Divergence tend to perform better over the weekend and convergence tend to do worse. Is that also your observation? I also want to ask you if you have any way to run "what-IF" scenarios across the portfolios? By this, I mean not analyzing trading stats but adjusting the actions taken to see the overall effect on the outcome of performance. Example: What-IF I let some select trades run over the weekend for past 3 weekends? I did some every extensive painstaking work in this space last year and could not come to any decisive conclusions (all in excel). I have now made up my mind to close all on Fridays but only from the standpoint of "mental peace during weekends" emotion and not based on any objective analysis. Hence seeking some ideas if I should continue this automatic and forced closure or re-assess. After all, there is no effort involved in taking realized profits (for me) and my drawdown manager function requires me to grow balance at a rapid pace to assist me in rapid equity growth and recovery. This approach really give a "feel good" feeling through out the week and feel bad feeling on Friday and feel good feeling on weekends. I am such a emotional mess LMAO

You spoke my mind bro. One day, I might publish a book on my journey and the lessons I have learned in past 3 yrs of developing automated strategy development :- It was an incredible eye opener to all the things I had to detox myself going thru this process. You are spot on with your comments on optimization and curve fitting. My current strategy simple cannot be backtested or optimized as there are just no tools available out there for me to use. It has over 350 variables that I have to set on every tick for me to control my exposure. As I have said before, my research tells me there are over 450K market impacting variables that could affect the closing price of one minute candle across 28 pairs and timeframes and sessions and other variability that market presents. It is just not possible to curve fit all these with any form of automation (but next progressive step for me is big data, machine learning and AI and I have some features already built in but mql really sucks in this space). If there are so many factors that can affect price, the only solution for me was to simply ignore price the moment after taking my entries....I have also completely remove wicks or HL from my entries and yet be able to pick tops and bottoms with decent accuracy.... a revolutionary concept that will only prove the test of time :-)

PS V....as a personal favour...
Sorry man. I just read your footnote. Perhaps you added this later and I am glad I caught it. I understand your concern. I will try to limit my deletes as it gives me the avenue to share my thoughts in an unrestricted manner to convey or contradict without leaving a permanent footprint on the internet. Regardless, I am at the end of my journey here on FF as it is time to stay focused on execution. i.e. make it big or break it & quit the trading for good.

Yes mate. You have two evils in which you can class every type of strategy....and this is how they respond to risk, You have either A) convergent or B) divergent conditions. This leads to up or down price movements at the granular level where price bounces between non-stationery equilibria and the only form of movement we as traders can turn into $. There are derivatives such as options of course using concepts such as straddles etc. for those who play the horizontal move...but in general....we are after either A or B.

They are symmetrical outcomes under normal *fairly* efficient markets where both deliver an equivalent outcome in terms of successful application..... but differ in the way this is achieved. Where they differ however is in extreme risk conditions where A) fails and B) thrives in conditions when markets are not efficient.

Divergent strategies are typified by the trend following enthusiast and have the following features:
1. Low win rate but high reward to risk;
2. Typically possess positive skew because of point 1);
3) Apply the philosophy of David Riccardo of cutting profits short and letting profits run;
4) Are open ended conditions as they can extend for any length of price duration away from a historical mean. Take for instance the Turkish Lira to see what I mean. You need to have your systems open ended to capture this uncertain extent of divergence. As a result this is why your reward to risk is potentially infinite with these models.

The mirror image to this is the convergent strategy which is typified by trading styles such as mean reversion or counter-trend trading has the following typical features (note the 'typical' being used as a generalisation):
1) High win rate but low reward to risk;
2) Typically possess negative skew because of point 1);
3) Apply the principle of buy the dips and sell the tips;
4) Have finite price move extent because they are travelling towards a historic mean....and when that is achieved they bounce back into a new autocorrelated phase of mean reversion to a non-stationery more recent historic mean or go through it into a new bearish divergent phase etc. These conditions have finite duration and you need profit targets as opposed to open ended profit targets to catch them. As a consequence this is why the reward to risk is lower than divergent models.

Now the key thing is that both trading philosophies require autocorrelation in the signals to be successful......but their risk signatures are different.

Think of a convergent strategy as being a rubber band that pulls back toward equilibrium. So we enter a trade when we have experienced an adverse move away from the mean with the intention that price reverts back to this mean where we take our profits. When it reaches equilibrium you would think that risk is released....well it is not. The market holds intrinsic risk in it's complex nature. It forever moves to new false 'equilibrium's. These equilibrium points are temporary thresholds that remain constant for a period of time while conditions persists. While markets tend to spend much of their time hovering around these equilibrium thresholds....this is when mean reversion shines and your trade sequence is typically small win....small win...small win...small win......but when a new threshold comes into effect....that is when the loss occurs and it usually is larger than those small wins......hence the convergent risk signature described before. If the market dislocation is extreme however the sequence of divergent losses have a more pronounced impact on the convergent trader with large loss....large loss....large loss sequences. This is when you realise that their was more risk in the system than you were led to believe.

Change the market state in a material way and you have a 'blow up' on your hands of convergent systems. This is because the market quickly adapts to new equilibrium thresholds. Convergent travels back to a past threshold....divergent travels forward to a new threshold. All convergent systems with their high win rate have more a Martingale risk profile built into them. It is a symptom of refusing to accept total losses to blow off risk steam.

Now think of divergent strategies as being the rubber band that snaps and prices continue to go on in the same direction to the new threshold. Divergent strategies however tend to have full risk release as the 'cut losses short principle' get's you out of the market quickly....and while you are out.....you hold no risk. As a result the equity curve of divergent strategies may be more volatile but they fully reveal the total risk of the trade....unlike convergent strategies.

Divergent traders are hoping for market conditions to transition between thresholds....and not stay on the same threshold. They therefore configure their trades to stand the test of time while markets are stationery. As a result they take many small losses during stationery periods in the hopes that one day...they will get the large gain to pay for 'em all when the rubber band snaps and a new equilibrium is created.

Now the divergent trader does well when markets are trending away from these past thresholds to new thresholds and that is when your sequence is large win....large win....large win....and you quickly recover from the drawdowns associated with stationery conditions of small loss.....small loss....small loss etc. as you are applying Riccardo's principle.

So....there is a sting in the tail for convergent strategies and 'why I keep away from them if possible'.....but I have been known to get a little pregnant with them....but never totally pregnant.

The third type of market condition that kills both strats is 'noise' which means that the frequency of autocorrelation in signal between divergent or convergent conditions is less sequential and more a mix of both and the result is that the sequence of conditions between these two conditions flips constantly. In noisy markets....both types of system fail as the noise is the result of a fusion of both convergent and divergent signals.

Both divergent and convergent strategies need consecutive autocorrelation of either divergent or convergent market conditions to beat the noise of the mixture. I hope that makes sense.


To the pro's and the guys who play around with diversified portfolio's....the concept of hard stops is a misnomer. It is all about applying performance based exists as opposed to hard stops or mechanisms that attempt to apply a fiat rule to risk management. By introducing hard stops into your strategy....9 times out of 10 you are deliberately introducing drawdown into your strategy which is the difference between the profit resulting from no stops applied versus the profit arising from stops applied.....and this is perhaps the most fatal floor that kills off any hope of an edge....and why everyone and their dog are seeking curve fit results that cannot stand the test of market uncertainty.

I will shortly be putting up a post to demonstrate the why? of this with solid support analysis as a foundation so every reader can dig a bit deeper into the reason behind such a bold statement from a supposed conservative risk manager such as myself.

So it is performance exits we are really after not arbitrary tolerance thresholds. You apply the arbitrary thresholds from the MAR ratio or other risk-weighted metrics at the level of the entire portfolio....not it's composite parts which you want to be simple and possess limited constraints or higher degrees of freedom to let the market and divergence/convergence do it's thing.

Given this basis of assumption....I follow this principle through by avoiding 'anecdotal risk statements such as the statement of ..."I only trade on thurdays'....or I only trade market opens....or I don't hold my trades over the weekend etc. These statements only have validity if your data analysis concludes this result.

The reason I am not crazy in doing so...by leaving my systems potentially open ended with their total risk exposure in an adverse event....such as a missile strike on Syria etc.....is that my diluted position size on every position on my portfolio means that the consequences are minimal that would otherwise blow the account of a non-diversified trader. Furthermore when the shit does hit the fan....stops are useless anyway as slippage flies straight through them....so it is more a feeling of being warm and fuzzy as opposed to an objective conclusion resulting from your data set.

Now I prefer divergent strategies to convergent strategies because of the following reason. With an asymmetrical stance towards profit taking whereby I have performance exits as an exit mechanism when things go bad and an open ended profit condition when things go good.....being a divergent trader under Riccardo's philosophy I can capitalise on the positive outlier. It is just the painful negative outlier that pin pricks me. The positive outlier can be a white swan event for me which may indeed be and frequently is on the weekend. These are the bread and butter that make divergent strategies shine. They capitalise on crisis alpha moments.

Nearly all the pro FM's in the CTA space apply this principle.....don't listen to news and just apply the same dare I say it V...."Universal" divergent recipe 24/7.

Bottom line is this. If you have done your backesting correctly and over a sufficient range of market conditions....then these tests include most of the the possible adverse conditions including weekend gaps etc......so if the backtests are still robust and demonstrate excellent risk-weighted metrics.....why would you change the way you trade under live trading?


Spot on V....hence my explanation in the preceding paragraph.


Definitely mate. As part of robustness testing you need to throw a range of 'what if' scenarios to your strategy. Once you have your benchmarks established in the backtesting environment and converted in Excel analysis....then you use random number generators with different degrees of extreme auto-correlation to test the efficacy of your systems robustness in a range of extreme conditions. You need to isolate your system weakness and understand the conditions when these occur. That way you can then monitor live market conditions from price data and identify when these instances occur....and how you can then adapt to respond to them.

Bing and bang...exactly mate. That's how I see it to.

great post, great mind. having questions and toughts like this (with syria eg. or what about this vulcano eruption with will affect comercial airlines or is best to buy stocks in glass companies and/or building companies after an earthquake or just buy the main indicies from that country?) when someone is building his system/approach is one of the best thing he can make because in short therm he will be aware that the system should manage abnormal events (what if one day oil countries will ask for euros, or jpy? or what if one day usd will be nio more reserve currency?)
everyone is asking what if gold&silver will be x usd or y usd but for mining them you need oil, then what about oil prices?

from soros:
Perceptions affect prices and prices affect perceptions
I believe that market prices are always wrong in the sense that they present a biased view of the future. But distortion works in both directions: not only do market participants operate with a bias, but their bias can also influence the course of events.
For instance, the stock market is generally believed to anticipate recessions, it would be more correct to say that it can help to precipitate them. Thus I replace the assertion that markets are always right with two others: I) Markets are always biased in one direction or another; II) Markets can influence the events that they anticipate.

As long as the bias is self-reinforcing, expectations rise even faster than stock prices.
Nowhere is the role of expectations more clearly visible than in financial markets. Buy and sell decisions are based on expectations about future prices, and future prices, in turn are contingent on present buy and sell decisions.
Once a trend is established it tends to persist and to run its full course.

Exactly M. I had to giggle because when I was writing about unforeseeable events I was thinking about Pip's quote on black swan events. Here I was talking about outliers when in fact the Black Swan event is when you are thinking in terms of currencies and you end up trading fish...... LOL. Totally unpredictable and unforeseen.....and it happens so frequently in financial markets as they are the derivative of the entire global economy and we are just trading representations or derivatives of that chaotic global system.

It makes you pause for a moment and wonder whether your system is really robust....or in the terms of a famous Aussie here from the film 'The Castle'.....tell im' you're dreamin.

Cheers mate

C

maybe you are aware or not: even if you have 2 system with same parameters/logic and just a very small difference in dd, return eg. and the only big difference being if you close position before weekend in one system and holding them in the other one is a subjectiv choice using system 1 or system 2 even if you've done objective backtesting.
markets like life have a lot of relativity in them

lol. yes mate, that's why i like markets so much. markets are so beautiful, like a beautiful women they can be volatile and tricky

years ago i used to play chess, i was good winning a lot of games just playing for fun and acording to pices like queen, bishop a greater value than for powns
one day also playing for fun i decided what if i will treat all pieces equally, in fact after 2 games i start played withouth my queen and i won 11-0 in that day. after some games i jocked with my friend telling him: look, we guys don't care about girls that are like this powns (ugly ones, ugly beaing subjective) but if after some square they can become queen, after some years some girls will be like models.

i approach market same way, i treath all postions like powns, after some time one of it will be my queen and will reward me greatly.
i see that you, ramadas and veefx are very smart guys and great statisticians and i really recomand all of you to read The alchemy of finance and Confusion des confusiones. take care guys

Cheers M

I will get my hands on your book recommendation. Sounds like it is right down my alley. Have a great weekend mate. :-)

Yep, Martin Lueck from Aspect Capital calls it agnostic approach. No bias whatsoever to position sizing, market choice based on recent history. More here:
https://www.toptradersunplugged.com/...spect-capital/



Why listen to him? From below chart it looks like he knows what he is doing, has rich experience, manages about 700 millions of dollars and was part of legendary AHL:

It is only click away, since it is public domain: https://babel.hathitrust.org/cgi/pt?...iew=1up;seq=17

You are welcome.

Thanks R. That saved me a trip to the mighty Amazon jungle :-)

Thanks for this R

It is interesting hearing Marty recount the 2014 year. Started with a horrible month in January and ended up supremely by December with a record for the fund. As he is talking about 2014 and other years in general I am referring to his performance record to get a feel for his psyche and how a stalwart approach to playing the market has really paid off :-)

Look how horrible Feb 2018 was for them.

Really ugly mate. What was given to them in Jan was taken away with gusto in Feb. It was a real rock and roller ride with the medium to longer term strats from the FM report we do. Fortunately the shorter timeframe strats such as our EDTT were spared the carnage....but that doesn't mean we are out of the woods :-)