Diversified Trend Trading Approach

And music:

Since you liked USD MXN, please see also USD TRY or USD SEK or USD NOK/.

Nice correlation table:
http://www.myfxbook.com/forex-market/correlation

Columns are editable if you are logged in.

Hey R

Thanks for sharing the system. Mate, I have a long list of projects pretty well defined that are all connected with my current portfolio approach that are steering me down a particular path that I need to follow. As a result, I probably can't do justice to testing your system. I hope you don't mind R. If the situation changes and I have a chance to test it, I will PM you the results.

PS After having a read of the system it makes good sense as a breakout technique adopting a multiple timeframe confirmation signal. However my path now is definitely a fully automated one with no discretion.....so that is the only point of departure for me from what appears to be a solid system.

C

Thanks for the tips R :-) C

Treasury-Bund Spread Gives Early Warning of the End

http://www.mcoscillator.com/learning...ng_of_the_end/

I understand.

How do you determine, that there is change in the direction of the SDC? Is there any rule that you start drawing new SDC in another direction? Sometimes I start drawing SDC short after price broke lower band in uptrend SDC, but then price start to make new highs and SDC long is still intact....

Cheers C. Very interesting.

While I try to avoid any form of prediction, it's hard not to form a similar view when I look at the performance metrics of the CTA's.

What your article and others, including my own thoughts below, tells me is that trend following is not dead at all. We are just in a pronounced mean reversion phase that is likely to end very soon in a snap. It appears IMO that the market conditions post GFC have led us into a trap thinking that that the markets have fundamentally changed...whereas I am of the view that the altered character is actually directly attributable to central bank intervention (Quantitative Easing) that has made all markets far more correlated than they were previously. As a result we simply have witnessed conditions where market conditions pre-GFC have now separated into periods of gross market intervention (globally correlated markets) periodically inter dispersed with periods of market readjustment (crisis alpha periods of equilibration).

The rise and rise of the S&P500TR post GFC is not directly related to traditional economic fundamentals but rather a symptom of 'loose money' from QE.......where the equity market has been a beneficiary of significant fund inflows simply because there are limited investment opportunities elsewhere.

As we all know markets have the ultimate last say and while they can be manipulated by interventionist policy for periods of time, they always have the last laugh with a classic 'snap'.

For example, have a look at the performance metrics of the Trend Following Index I track, which is a composite of 54 diversified trend following funds that have been in operation since Jan 2000. Now there are ways you can actually make sense of market conditions without falling into the trap of simply using historical charts that can be quite misleading in nature. The way I do it is through forensically examining equity curves.

An equity curve plots how a chosen system or broad class of systems, if using an index of these systems, responds to market conditions. There are two broad market conditions on any liquid market....namely convergent behaviour (mean reversion) or divergent behaviour (momentum/trend following). Using an index of trend following funds we can then see the wood for the trees and discern market character in terms of periods of divergent behaviour and periods of convergent behaviour. So now have a look at the equity curve of the TF Index below and see what it is saying.



Things to note:

1. Despite the calls that trend following is dead.....The performance returns since Jan 2000 have actually been consistently delivering approximately 9.3% PA (refer to the regression line acting as a line of best fit across the monthly returns of the TF Index). What this is telling me is that the market actually has not changed much at all....but rather it appears that it has changed as we now have distinct phases of mean reversion versus momentum.

2. While the overall returns are still consistently delivering superior returns to the S&P500TR Index, what we are finding is that the markets have clearly separated into periods of protracted mean reversion inter-dispersed with periods of acute crisis where trend following shines. These market characteristics appear to be directly related to central bank intervention where the mean reversion phases are clearly coinciding with quantitative easing measures post market shocks (as seen by the periods of significant trending).

3. The equity curves produced by the TF Index are representative of all market sectors given the diversified natures of these FM's. The clear separation into trending and mean reversion periods is symptomatic of the fact that when markets are highly correlated, benefits of market diversification are lost and markets tend to be mean reverting in nature. Conversely when markets are un-correlated, benefits of market diversification are gained and this occurs during trending periods. What we clearly are witnessing is that all liquid markets right now are very highly correlated. Before 2008, markets were less correlated allowing for diversification benefits to be achieved by market diversification. Today, such benefits are not delivered from market diversification except during crisis alpha moments. That's why at the moment more consistent returns lie in system diversification as opposed to market diversification.

4. If performance returns of the TF Index are to continue delivering consistent returns of 9.3% per annum (as an historical average), then it appears that we may be at the tail end of a mean reversion phase and that the next trending phase is likely to be of a similar or greater greater magnitude than 2014/2015.

5. The performance of the S&P500TR has been in one of the longest bull runs of history post GFC stimulated by QE. The questions you need to ask now are: What is holding it up and how long can this continue? What is happening to CB interventionist policy now and what does that mean for the future?

6. Look at how the market conditions between trending and mean reversion phases are actually rhythmically beating along the line of best fit. Look at where we are now in relation to that line of best fit. Do you see what I see?




C

Thanks R. I appreciate it :-)

I am always one step in the past referring to my previously drawn SDC and where current price lies in the context of that trend for that particular timeframe.

For example have a look at the Daily chart below. Note that this trend has been a long one and we have narrowed our SDC to 0.5 when we reached a profit target of 2.5 x R. In the context of that chart we have been enjoying a great long trend and currently price is still a long way away from the trailing stop. The trend remains alive so long as conditions of drawing the SDC are not breached. If price continues to go up, we re-draw the SDC every time a new high is achieved. This continuously adjusts the placement of the trailing stop with the new price data.

If no new high is reached and price now wanders down to the trailing stop we don't make any further adjustments to the SDC. Unless a new high is reached, the SDC is locked in and defines our current trend.

When the trail (projection) is hit we are out of the trend. This is because in accordance with our definition of trend, the long trend on D1 is now over. We now choose the prior swing high as the new commencement point for a short trend and redraw a new SDC from that prior swing high in the short direction.

Now if price then resumes back in the direction of the prior trend we get a trend end on the short term short trend and we are left in no mans land waiting for a new trend definition to evolve. If a new high is made, then redraw the prior long trend SDC to capture this data. The new amended SDC will widen to encapsulate the more recent retracement move and we are now ready to trade long again with a new breakout with a revised SDC and trend definition. If a new low is made then we redraw the short trend SDC to capture this new price info.....and so on and so forth.

From this process you can see that the important info that is critical to this technique is how we interpret what is a swing low and swing high. This determines where we commence our SDC draw from and defines what we interpret as the trend.



I hope this makes sense R. Let me know if you have any problems with this interpretation.

C

OK, thank you. I was wondering when you start drawing new trend and now is clear. After first new opposite low/high outside of trailing projection.

I am sure you know about fractals as a way to define high/low. Did you do any testing on it?

http://www.investopedia.com/articles...6/fractals.asp

Yep R. Significant effort in terms of the application of fractal analysis has been undertaken already in the vain attempt to fully automate this approach with limited success. It's not that I and another on this forum haven't tried ...it's just that .....dare I say it....in terms of selecting the appropriate swing high and swing low takes a number of factors into account such as trend quality, momentum etc.....and this is where full automation sucks and the powers of discretion and logic have an edge. See R....you told me so.

I am hoping that one day I can get some good info from Bill Dreiss on how he fully automates his successful approach apart from info that has already been made public. One of my earlier posts in this thread has some good PDF's from Dreiss Research on the broad methodological approach they take which gives some clues. :-)

C

EDTT is most promising according to current results...

USD JPY: Not really sure if I should draw new baby uptrend or wait till old uptrend will be continued?

I put prediction lines in green. Am I correct, that we wait till these prediction lines will be broken? Nothing to do till then?

Also time frames multiples between d1 and h1 and h1 and m30 are bugging me. M30 and H1 is basically the same chart with multiple of 2 between them. 1d and 1h are multiple of 24.

Why not h4 and m30 for secondary trades? - multiple of 8? H4 was proven as best time frame with least signal to noise ratio and optimal to trading here: https://www.forexfactory.com/showthread.php?t=336715

And M15 is too unpractical to maintain 24 instruments? Or too much noise? Or trends are short on this time frame?

Yep R. You are getting it. That's a no mans land example for USDJPY for M30. I would be waiting till confirmation of old trend and then redraw old SDC.

With regards to pending entries....I personally use these to coordinate trades between timeframes so I progressively scale into trends as I move up from M30 to H1 then H4 then D1 then W1 etc.

Given you are working on M30 in the example above your pending entries are fine....however if it was me I tend to stretch them further into blue sky to avoid fake-outs. In the last few years with mean reversion in force, placing breakout entries a few pips away from a previous high/low has tended to end in tears for me.

The pending entries are used to:

1. Co-ordinate entries across timeframes so you are never too exposed with trade risk. For example if I enter on M30, then my pending entry for H1 is placed at a position where the worst case scenario for the prior M30 trade is above breakeven with regards to the position of the trailing stop;
2. Avoid over-trading by placing pending entries immediately after a closed trade in zones where price really has to stretch to achieve this new pending order. If I therefore enter a sideways congestion phase, I never get suckered into these with multiple trades;
3. In times of building drawdown I progressively stretch pending entries further away from current price making price have to work harder to trigger new entries. This reduces the effect of unfavourable drawdown periods and preserves capital. When volatility returns, the pending orders then progressively ratchet down closer to price.

C

It's up to you R as you need to be happy with the logic of your approach. Just keep in mind that we use multiple timeframes as:

1. A method to scale into a trend adopting a bottom up approach to riding an emerging trend;
2. A method to capture swing moves (fractals) on multiple timeframes;
3. A portfolio diversification measure which means that it is preferable (if possible) to have different SDC's for each timeframe that are relevant to that timeframe. If that is not possible given the nature of current price action then simply use H1 for the purposes of identifying a future entry to scale into the same trend as M30.

M30, H1, H4, D1 and W1 work for me....but there is no reason why different configurations shouldn't work.

With regards to shorter term timeframes such as M15 etc......the reason I don't trade these is predominantly due to the practicalities of trading a diverse portfolio. I hate being an octopus especially when things are moving fast.

Cheers

C

Thank you!

Bye bye GBP USD, we can try to enter again....
But indexes say, GBP will fall at least to mean.





From indexes we can see GBP JPY short would be nice and EUR USD short, since GBP and EUR are topping and JPY bottoming.

Still riding crude:

And gold is getting frisky:

Do you take into account in EDDI, where price is in larger frame channel? In some analysis you said overbought/oversold. Was my entry correct (near lower band on h4) or should I have waited till price was over mean (for downtrend) on larger frame?

I am sorry if I ask too many questions...

Could you please tell me how do you draw channels? Short or long?

Totally agree with you as long as we have easy Money market benefits, I have question from your point # 3 The equity curves produced by the TF Index are representative of all market sectors given the diversified natures of these FM's
Is there a common sector or multiple sectors from 54 funds that is cause of mean reversion? Are these sectors showing sign of bottoming?

Traveling Wilburys - End Of The Line


Jimmy Cliff - I Can See Clearly Now (with lyrics) HD

Evening R

The EDDT is a breakout technique and the SDC is used differently to the way it is applied with the DTTR (the retracement entry technique). With breakout techniques the info I am after from the SDC is the trend momentum, direction and volatility. I glean this from an SDC set at 1 x std deviation.

With the DTTR I am after 'precise' trend entry placements in areas offering excellent R:R characteristics if the trend is valid. The way I draw the SDC with the DTTR is different to the EDTT and it's aim is to define the characteristics of the primary trend on the weekly timeframe and then drill down into D1 to find optimal places to enter at key retracement zones. I encapsulate all 'adverse' volatility price action within the SDC from the swing high or swing low on the weekly timeframe to define the volatility range of the trend. Once this is done I can then plot retracement zones which offer excellent entry points with relatively high position sizing to hopefully ride some powerful moves.

EDTT

Firstly with regards to EDTT and your EURCAD example, this is how I currently have EURCAD configured on the 6 timeframes I track.



I will step through my logic.

Monthly Chart
Starting with MN. My SDC currently on the MN chart suggests that the primary trend is currently changing direction from long to short. Note how the pending orders are placed a long way away from current price. To ride the trend on MN, we require price to move considerable from its current position either long or short to provide sufficient monthly bars to suggest a trade-able trend is on in that timeframe.

Weekly Chart
Note how that as we get more granular as we step down to a lower timeframe there are more bars available in the short move to better reveal what is going on with the recent trend change. If I apply an SDC to the most recent swing high on the weekly, there still is insufficient data for me to conclude that we have a short trend of substance on the Weekly timeframe. Also note where my pending orders are on the weekly and compare their placement against the pending orders on the monthly. You will see that if we enter a short trend when the short order entry is activated on the weekly, we still need further for the trend to develop before we add another short position on the Monthly timeframe. This is how I stagger my entries on each timeframe. Ideally you do not enter another timeframe until you have at least achieved breakeven on the other timeframes in relation to the trailing stop once the trade is active.

Daily Chart
Still not enough price data to get excited with a short trade. We need price to stretch on D1 to get an itchy trigger finger. Note where my pending order entries are with respect to the W1 and MN charts.

Clearly, given the recent trend change on MN, all the action is reserved for the shorter timeframes to hopefully catch a ride on an emerging trend.

Let's start on M30 and then move on up to H4.

M30 Chart
The pending order short and long was placed when I was knocked out of the long trend on 28 June and I haven't moved it since. At that point in time I placed my short entry pending order at a point in clear blue sky short (with breathing room) below the prior swing low on about 20 June. There were no other trades active at the time on EURCAD, so I placed pending orders on all timeframes at this point in time to allow for a sequential order entry into a nice trend either short or long.

Price is getting closer to the pending order short and if activated, we would be short on the M30 timeframe with a possible nice early entry onto a longer term short trend across all timeframes.

H1 Chart
Unfortunately there is not much difference between the price action on M30 and H1 and as a result the SDC for both M30 and H1 commence at the same swing high of 28 June 2017. However note where the pending order short is placed on H1. For this trade to be activated I will already be beyond breakeven on M30.

H4 Chart
My SDC on H4 commences from the swing high on 2 June 2017 and therefore the SDC has different characteristics on H4 as that of H1 and M30. This is ideally what we want to achieve timeframe diversification. Note where the pending orders are on H4.

Now when you look at all pending orders and SDC's in context you can see that you have captured price action on all swing moves and potential trends that are available. When volatility heats up and some great trends emerge you can find that you have 6 positions running on all the timeframes and your profitable soars while at all times you have been managing your risk and trade exposure.

I hope this helps mate.

Now onto DTTR

DTTR
I did a bit of a write up about the DTTR on this post so I want waffle on about it. It is a very different approach but uses the SDC (as I love the SDC) to assist in interpreting trends. It's probably best to stick with breakout techniques first before getting into the DTTR....however the two techniques when combined really assist in smoothing the portfolio equity curve. During periods of mean reversion, the DTTR is pretty powerful but during trending periods the EDTT cleans up. At the moment I only trade one timeframe on the DTTR yet down the track will be looking at possibly adding further timeframes like the EDTT.

Here is how I have plotted the EURCAD on the DTTR. Price is currently testing the previous long trend and we may find a nice short trend emerging from here on. At the moment I am in no-mans land waiting for a confirmed entry. If it goes long and touches the pending entry line I go long and set my stop and trail on the pending trail line. If it continues to go short I wait for a new retrace once the trend break is confirmed to enter short.




Cheers R

C