Diversified Trend Trading Approach

Yes, I agree with NorthTrader's and 21vs7's comments and really appreciate what you are doing here C.

It's a pleasure K. I hope all is well mate in your beautiful part of the planet.

May Interim Performance Summary - Discretionary Systems - Breakout Techniques (DTT, EDTT, EDTT1)

Man O' Man....this is getting painful. We are officially experiencing the 'doldrums' with this technique. May is proving to be a problematic month trying to catch the trends and obviously mean reversion is ruling the day. What started as "All Quiet on the Western Front" is now turning into a crushing defeat.....so it's time to get defensive.

Ok lads.....let's dig a bit deeper. Looks like we are in it for the long haul.



As a discretionary trader, this is where you have flexibility on your side by making a few simple decisions to simply slow down the bleeding.


Statistical Summary of Closed Positions for the period 1 May 2016 to 21 May 2016

Overview
May is proving to be a difficult month on the trend front. We are down about 1% for the month and there is not sufficient unrealised profit in the portfolio to maintain an aggressive stance. Drawdowns are starting to accumulate and to prevent over trading, risk management tactics are now being deployed such as deliberately reducing trade frequency by making price have to stretch further prior to an entry through more distal pending order placement. Such measures still allow you to participate in directional price movement when it occurs, however it reduces the propensity to trade when the market is clearly in a mean reversion regime. As mentioned last week, in my opinion, it is these periods that define your ultimate trading success with discretionary methods as capital protection is probably the only element of trading that we can directly control.




Top Closed Trade so far for the Month




Current Open Positions





What's Trending at the Moment According to this Technique?

AUDUSD
Trading on 5 of the 6 timeframes and almost fully loaded.




USDCHF
Has started well on M30 and have recently opened up a trade on H1. If the trend continues we will progressively become fully loaded. So far so good.....but I have said that before.



That's it for the breakout discretionary systems. Full details are contained in the attached spreadsheet.

My next post will feature the interim results of the Discretionary Retracement Entry approach (DTTR)

C

PS OMG what have I been thinking. How could I forget the essential music!!!!

May Interim Performance Summary - Discretionary Systems - Retracement Entry Techniques (DTTR)

Below are the interim results for the recent entry into the discretionary systems currently being trialled. If results prove to be sufficient over a sufficient data sample, then both the breakout and retracement approaches will be consolidated under a single portfolio umbrella.




Statistical Summary of Closed Positions for the period 1 May 2016 to 21 May 2016

Overview
After an excellent early start, things are starting to slow down on the mean reversion trend front with more subdued volatility. A small loss for the month of <0.5% has been incurred which is immaterial given the extent of unrealised profits currently sitting in the portfolio. Fortunately this technique takes advantage of mean reversion back into the underlying trend which, given current market conditions, is giving an edge to this technique over the discretionary breakout techniques of the DTT, EDTT1 and EDTT. As a result, this discretionary system is likely to be a worthy addition to the arsenal that under a system blend with the breakout systems, is anticipated to significantly flatten the equity curve of a blended system especially when the market enters prolonged periods of mean reversion.

Despite the more subdued volatility, the unrealised profit of the portfolio will assist in progressively releasing positive cashflow if the market remains in the doldrums. Please note that I have included the result of the AUDCAD trade from April as this was the only trade exit made during the month and represents a profit take on half the initial position made at 3xR according to this technique. Note also that this technique only trades according to the D1 timeframe. At a later date, dependent on the results we get, I may start opening up additional timeframes to trade the DTTR.




Top Trade so far for the Month




Current Open Positions





What's Trending at the Moment According to this Technique?

AUDCAD - refer to Top Trade of the Month (above)

AUDJPY


CADJPY


DJI


AUDUSD


GBPSGD



Another lackluster week. C'mon trends....where the bloody hell are ya :-)

Aveagreatweekend.

C

Worth The Wait....... Part 2

Following on from post 922 the discussion with Neils continues with Robert Osario from Dunn Capital, Alex Greyserman (ISAM) and Kathryn Kaminski (Campbell & Co) who delve further into the Simplicity of Trend Following. We start getting into the meat and bones of trend following in this podcast so enjoy. Just remember to ensure you listen to the earlier podcast session so you can catch up with the discussion.

Here is a summary of the nuts and bolts of the discussion with a bit of poetic licence I have applied to flesh out some of the points raised.

As a definition, trend following is simply a specific set of rules that defines your trading parameters. In it's simplest sense to be a trend follower, you assess whether price is up or down with respect to the average directionality of your chosen look-back period and then enter a trade in the direction of that historic directionality and let profits run (while price movement obeys that condition) and immediately exit from your position when that condition is no longer met (cut losses short).

There are many permutations of how you play this game but in essence it is simply about defining rules to get in and get out of positions that exploit the propensity of your market to move in a directional manner.

Simplicity and robustness goes hand in hand. For example the simplest way of trading a diversified portfolio is to apply an equal $ allocation to each position (or a 1/n solution where n = the markets in your sample investment universe). What we therefore are ensuring is that all positions are treated equally and that no attempt has been made to curve fit the portfolio to the best recent performers. This way, we are ensuring that we do not adopt a predictive posture towards the future of the market. To adopt a view of the future, you need to have proof and this is a very subjective enterprise when dealing with complex efficient markets. We cannot make an informed guess about expected returns. The best we can do is make a bet based on past history.

You should not introduce complexity gratuitously. Be as simple as possible ...but not simpler. Ensure that your addition to your model is responding to a large data set and not overly skewed (biased) by small data sets such as recent market conditions. We evolve over time to survive. As markets adapt we compete and those that don't compete don't survive. So it is essential that our models utilise extensive data samples so that we can detect the difference between seasonal (short term) changes versus fundamental changes.

For example you will find in the CTA space that there appears to be a growing dispersion in returns of the trend following FM's. Some have posed the suggestion that trend following systems are becoming less correlated, but this is not actually the case. The reason for that significant dispersion is due to different firms including additional 'non-trend following' convergent strategies such as mean reversion, counter-trend, pattern recognition systems in their portfolio to respond to current mean reverting market conditions. Underlying their portfolio's however are the dominant trend following systems that kick into gear when we experience regime changes. However even for those traditional trend following firms, return dispersion can significantly differ.

If you only consider the dedicated trend CTA's that do not stray into alternative 'non-trend following strategies', you will find that their correlation is very similar at approximately 0.7. (Copernicus note: For example refer to the table below where I have highlighted in green all those CTA's who possess correlations of >0.65 with the TF Index. These are the funds that employ traditional trend following systems. Those that possess significantly lower correlations are straying into alternative systems to further diversify their portfolios and flatten their return profiles).

Now most make the mistake that with with such high correlations, the returns of the funds are going to be similar however unless the correlations are really high (say approaching 1.00) then we still get significant dispersion in the results. So do not use correlation as a definitive guide to assess performance returns.



Refer to the paper attached below relating to return Dispersion of CTA's.

When everyone is following similar trends, the returns still vary extensively as it depends on when they enter, when they exit and the position sizes adopted. So given this intrinsic feature of return dispersion, THAT IS WHY WE DIVERSIFY to flatten the volatility of return profiles. Diversification is not simply about diversifying across different markets. It is any range of methods used to reduce the volatility of returns such as market diversification, data stream diversification, timeframe diversification and probably most importantly system diversification.

Diversification into different markets has never been easier. There is no excuse not to use it today. It is the only free lunch........and portfolio managers who do not include trend following in their portfolio's are simply ignorant and doing a disservice to their investors. If you are a trend follower, a focus on a single market will get you nowhere. Trend following and diversification must be melded together as diversification is a major mechanism in which trend followers control risk.

The heart of trend following is:
1. Generation of signals;
2. Position Sizing;
3. Exit strategy; and
4. Overall risk management.

All trend followers do is define rules that respond to typical questions that every trader faces. When to get in, when to get out and what position size to adopt however the entry point is not the most important consideration. Modern trend following requires you to manage risk at all times of the trade. Precision of entry is simply not that important. We are simply risk managers. In order of importance:
1. Exits; then
2. Position management; then
3. Entries.

In determining the most robust parameters, we use meta parameters that determine the choice of specific parameters to use. We do not want to use a set of parameters that have performed the best over a chosen lookback as it almost surely won't be the best performer in the future. It will however will give some indication and hence we use metaparameters to define a broader set of parameters such as a correlation or standard deviation rule. If you follow a pure scientific process you first develop a hypothesis and then you test that hypothesis for parameter stability. It is not a question about first picking parameters that best respond to current market conditions but rather following a systematic scientific approach that validates a view (hypothesis) as opposed to mere data mining.

I had to laugh when this example was stated by Greyserman (the bronx voice)...... "a trend follower approached me and stated he obtained a Sharpe ratio of 2.0 so I asked him how did he get that if he applied his approach over a diversified portfolio of say 100 markets......Well he said ....I only used the best 30 of the markets in my model..... which immediately identifies that he curve fitted the approach using hindsight as a measure in his selection."

The key to robustness is not to fall into the trap associated with only using a small data set. For example bonds have performed well over the past 30 years but this of course has not always been the case. To make a decision that you will bias your portfolio with bonds therefore is a curve fit solution. Avoid any form of optimising bias where possible, and use a logical hypothesis as a basis for your decisions that has been as strenuously tested for it's validity as possible. If you don't have actual data, consider using synthetic data to test your hypothesis.

If you are seeking to include trend following CTA's in your investment portfolio....you need to diversify across trend following CTA's....perhaps up to 3 or 4. More diversification tends to improve results....however as you increase further, the marginal utility of diversification declines.

......anyway, that about sums things up.

Enjoy the podcast.

C

Sunday is drawing to a close..... but a final last ditched effort to squeeze the last few drops out of the weekend.

Good trading for the week ahead. :-)

thanks for the tunes and info C.. have a good one.

Brilliant! Thank you very much for doing this, C.

No probs NT. I think you may be onto something in encouraging me to do these summaries as it is helping me on my journey by forcing me to slow down and think. :-)

May I join in to assist you in my bumbling way ? Excellent work(he's Copernicus what else do I expect) that serves as a great point of reference with good music for accompaniment.

Cheers S......the music is the important stuff.....the trading, well, that's just the fluff around it.

Cheers Ja. Another week kick starts and I must admit....I'm not in the mood for the grind. I think a holiday is in order.

I hear ya mate.... To be honest I could do with one too ..Friends of ours have got a place in the country.. its nice.. my wife and the kids are there from this Weekend on-wards.. I think Ill get on it...take what I can off work and join them..

Sounds like a plan mate. Life's too short. I think I might be doing the same thing and heading off to the snow in July with the wife and kids....to hell with the grind I say.

Luv'd that film clip mate. Some very talented people on this earth.

Question for people on this thread who trade trends:

Do you focus on 1 key strategy / approach to trading or do you trade multiple styles / strategies?

Ideally you are after benefits of diversification so in general, the more systems the merrier provided that each system contributes positive expectation. That way you can flatten the equity curve and make more efficient use of your capital using enhanced position sizing while minimizing drawdown at the same time. Even highly correlated systems under diversification assist in smoothing the volatility of returns but ideally, non-correlated systems create added benefit.

May Interim Performance Summary - Discretionary Systems - Breakout Techniques (DTT, EDTT, EDTT1)

May is digging us deep into our trenches and the defensive play is the order of the day. " Keep positive lads.....we'll lick these blighters soon enough and be back to Blighty before Christmas. Watch out for those damned drawdowns and don't shoot until you see the white's of their eyes. Now Watkins....did you bring the sandwiches?"





Statistical Summary of Closed Positions for the period 1 May 2016 to 28 May 2016

Overview
May is proving to be a difficult month on the trend front. There is not sufficient unrealised profit in the portfolio to maintain an aggressive stance. Drawdowns are starting to accumulate and to prevent over trading, risk management tactics are now being successfully deployed through more distal pending order placement to reduce trade frequency and stem the bleeding. The week has therefore seen only a marginal decline in realised profits and the drawdown is being effectively managed at approximately 1.1% of trade capital. As mentioned in the last few weeks, it is these periods that define your ultimate trading success with discretionary methods as capital protection is probably the only element of trading that we can directly control.






Top Closed Trade so far for the Month






Current Open Positions




What's Trending at the Moment According to this Technique?

AUDUSD
Some profit taking has occurred on some timeframes (H4 and M30) but H1 is still going strong with a bit of life re-surging in D1, W1 and MN. We are looking for re-entries back into H4 and M30 if the short trend continues.




EURUSD
Some life is coming back into EURUSD with short trade opportunities opening up.




USDCHF
Has started well on M30 and H1 with an imminent trade entry for H4 looking likely. If the trend continues we will progressively become fully loaded. So far so good.....but I have said that before. Go the Swiss cheese.






That's it for the breakout discretionary systems. Full details are contained in the attached spreadsheet.

My next post will feature the interim results of the Discretionary Retracement Entry approach (DTTR)

C

May Interim Performance Summary - Discretionary Systems - Retracement Entry Techniques (DTTR)

Below are the interim results for the recent entry into the discretionary systems currently being trialled. If results prove to be sufficient over a sufficient data sample, then both the breakout and retracement approaches will be consolidated under a single portfolio umbrella.





Statistical Summary of Closed Positions for the period 1 May 2016 to 28 May 2016

Overview
After an excellent early start, things are starting to slow down on the mean reversion trend front with more subdued volatility. The month so far has produced a break-even result with respect to closed trades, however a substantial unrealised profit from open positions is starting to accumulate. Fortunately this technique takes advantage of mean reversion back into the underlying trend which, given current market conditions, is giving a current edge to this technique over the discretionary breakout techniques of the DTT, EDTT1 and EDTT. As a result, this discretionary system is likely to be a worthy addition to the arsenal that under a system blend with the breakout systems, is anticipated to significantly flatten the equity curve of a blended system.

Despite the more subdued volatility, the unrealised profit of the portfolio will assist in progressively releasing positive cashflow if the market remains in the doldrums. Please note that I have included the result of the AUDCAD trade from April as this was the only trade exit made during the month and represents a profit take on half the initial position made at 3xR according to this technique. Note also that this technique only trades according to the D1 timeframe. At a later date, dependent on the results we get, I may start opening up additional timeframes to trade the DTTR.





Top Trade so far for the Month





Current Open Positions





What's Trending at the Moment According to this Technique?

AUDCAD


AUDJPY



GBPSGD
Refer to Top Trade on the Month above



CADJPY



AUDUSD




FTSE





It is tailing off into a very slow month guys....but keep the faith. Who knows what is just round the corner? :-)

Aveagreatweekend.

C

mr copernicus, I know the question I have for you is relative and can only be answered with years of experience and learning curve of each person, so in advance I apologize if I ask for the more possible technical answer.
Regarding draw (W1) SDC you raise to recognize the most obvious chart swing points but this is where the most subjective and difficult part of the process comes as the more look at the chart the more obvious points meet and more confusion have to draw the channels.
To help my self to visualize these points, I improvise two techniques but nevertheless were not enough to give me the confidence to implement your methodology.
The first was place a ZZ (24,5,3)
The second was to try (subjetive) to establish the very last SDC and as soon as the price breaks the upper/lower line, draw a new SDC to set the new trend and operate according to it until price break again some these levels and continue the process .
I wonder if you have any tecnical parameter (or about tecnical) to establish these swing points necessary to draw the channel and how deep in history should be necessary to established it.
Another question: it is important the chart zoom level in order to set the length of the channels?
Thank you and B Regards.

Hi jusuir

I understand where you are coming from, but unfortunately the determination of the swing high/swing low is not an accurate science and as yet, I haven't been able to fully automate the discretionary entry. The trend definition and determination of swing high/low to extend the SDC from is very subjective in nature, however perhaps if I give you the logic from a few examples, it may help you determine your own rules. You will see that the method I take is not one of 'accuracy' but one of making price action have to obey particular conditions before it tempts me to pull the trigger. You should define your own logic and rules that makes sense to you with the main aim of reducing your propensity to overtrade.

Here is a general guide to what I do for the DTTR techique (retracement entry), but recognise this is not an exact science. Despite the subjectivity, the intent for me is to decipher where price currently lies in the context of recent price action on W1. So when I come to a raw chart, I am always looking at the right edge of the chart and determining where price sits in the context of recent price history. It is deceptive if you come to a historical chart and simply identify swing highs/lows using hindsight as unfortunately this is not what you face when trading live.

So for example I will plot a few examples below and give my reasons for how I plot the price bias.

Below is a raw chart of EURUSD on W1 and I am interested in price action over the past year or so as a basis to attempt to decipher the overall direction of price during this period.




My zone of focus is highlighted by the beige rectangle below.



Now clearly in the context of W1 we have a clear sideways congestion range (well at least that is how I interpret it). You could actually conclude that it was trending sideways. There is no signal generated from this sideways congestion that makes me conclude that price is about to breakout from this sideways congestion into an up or down trending move, so I place breakout levels well above/below this congestion which price needs to punch through for me to consider revisiting this instrument.



The reason why the breakout lines are located well above/below this congestion is that I am ensuring that price needs to have to work hard to trigger my interest and we need a clear indication that upwards/downwards directionality is in force before we jump in. If we get price moving in a directional manner smartly to these defined lines of interest, then obviously momentum has picked up and we could have a nice trend emanating from prior congestion.

Anyway, I leave this instrument for the moment and move onto another and won't come back for a revisit until price get's close to those drawn lines.

Let's look at Gold

Here is my zone of focus. Now I need to consider this in the context of the history and clearly we have had a protracted bear phase for this instrument since 2011.



I now want to define where price currently sits in the context of this history. I can clearly define by eyeballing where the bear phase extended from so I define my swing high (green rectangle) and draw an SDC that fully encapsulates price movement from this swing high to the most recent swing low (blue rectangle).

As it turns out, to fully encapsulate price my SDC needs to have a width of 1.30.



Now I assess where price currently sits and note that it tested the outer limits of the SDC but clearly appears to have retraced back into this primary trend bear movement. Now I am excited as this represents a possible opportunity to catch a solid mean reversion move back into *my definition of a* trend.

So I now need to incorporate that recent price data into my SDC channel so I need to slightly widen it further to 1.45 and extend rays to ensure I have encapsulated all price history of this bear phase move.



Now I am ready to subdivide my SDC to identify my entry condition. AS the SDC = 1.45 I divide this by 2 to create an SDC of 0.725.

I am now waiting for price to retrace through the green highlighted line which is my confirmation signal for entry to signify that mean reversion back into trend is underway. Clearly if price decides to change direction and head back out of trend then I use the red line as my signifier that the trend I am currently riding is no longer valid and I immediately get out. It furthermore acts as a trailing stop.

Note that the SDC is dynamic and that when a lower low (short) or higher high (long) is made you redraw the SDC from the original starting point to that new bar. You also need to adjust the width to continuously ensure all price of your chosen history look back is contained within the SDC. This therefore ensures that your trend definition is always representative of current price action in the context of historical price action of your chosen trend.




....anyway, I hope this helps a bit. It is very subjective, but as you become more familiar with approaching trend trading this way, you start recognising that the entire game of trend trading is not an exact science. The important thing is that you take a small position based on your assessment of whether price is moving in a directional manner, define the parameters upon which your trend definition is based and the conditions that your definition breaks down, and provided that price moves in accord to those conditions, then let it run. If it doesn't then get out as soon as your definition breaks down.

Do this over a diversified portfolio of liquid instruments, treating each instrument exactly the same and it is as simple as that.

I hope this helps jusiur. :-)

PS The key to this trend trading game is not accuracy in entries. A bit of simple logic helps or at least gives confidence in decision making to help you pull the trigger. The major aim is for you to adopt a trading logic that is repeatable that simply seeks to follow overall price movement once you have pulled the trigger, as unfortunately once that occurs there is no going back. Even random entries do the trick provided you can develop a methodology to manage the trades in the same repeatable way, once you have entered. The reason however that I don't take random entries is that I define rules that restrict the propensity for overtrading. I therefore in my own roundabout way, create particular conditions in the setup phase that make price action have to work hard to get me interested.

They key is in letting profits run and cutting losses short. That is it....plain and simple. We can never predict where price will go but due to those fat tails on non-normal distributions, you will occasionally find you are in the right direction to capitalise on this due to a feature called autocorrelation (price action persistance) which is being driven my underlying fundamentals and manifested through participant behaviour. If you adopt the principal that you take an entry in the direction of recent directional action then it might just turn out that it continues on in that direction for a long time and with good momentum enough to stay within your trade risk parameters and you simply book those profits when they arise.

The critical part is to manage risk along the way and across the portfolio. Simple history reflects that a single instrument may only trend significantly 1-2 times a year. With that knowledge you can afford to be brutal in reducing trade frequency on a single instrument and entering late into a trend of significance. Trade a diversified portfolio and you can then be very selective in raising the thresholds for your entry criterion to simply avoid overtrading which is part of your risk management bag of tricks.


Cheers

C