Diversified Trend Trading Approach

It's not that I haven't tried to make this fully automated but the essential nature of this particular method is the discretionary decisions made in determining the context across multiple timeframes which is a necessary pre-requisite, and applying the logic of this context to when and where to enter on a particular timeframe. Upon entry however, it becomes fully mechanical as the SDC takes over in managing each trade to exit.

The MN and W1 timeframes are my indicators of the macro picture and my lower timeframes are where I place sniper entries in accordance with the broad market context.

Numbnuts in his great new thread has also inspired me to look at ways to incorporate retracement entries into my discretionary systems that potentially offer great R:R and can supplement the breakout nature of the discretionary strategies. While I am not using FA like COT charts in the process (as I am a newbie to FA), I am replacing the FA components of his technique with the statistics of the discretionary method working off longer term timeframes (MN and W1) to give definition to a trend and placing sniper retracement entries on the D1 timeframe at extremal retracement zones to capture swing components and high R:R entries for longer term trend following.

I will be discussing this in later threads when I have honed this additional approach which hopefully will be a valuable component to the suite of discretionary systems described here already.

But below is a sneak preview of an orderly trend I am riding using this alternate retracement entry method. I happened to start trialing this method in live forward testing commencing 18/4/2016 so unfortunately missed out on previous retracement levels on about 5 April 2016....however you have to start somewhere. The trade comprises two positions. One position closes on entering the lowest SDC zone (when short) of a secondary trend (or fractal) or at 5xR.....whichever comes first, whereas the remaining position rides the long term trend and exits on touching the outer SDC of the primary trend.

When you nail the retracement entry on a primary trend you can expect some glorious multiples of R.

New Discretionary Strategy Addition - DTTR (Diversified Trend Trading Retracement)

I thought I would provide screen shots of a new approach I am trialling as an addition to the suite of Discretionary Systems described in this thread. Without having to go into all the gory details of how this strategy works, they say a picture paints a thousand words, so I will let you all have a crack at breaking down the logic and technique. Here is a hint..........As a general principle, the technique defines overall price bias on the W1 timeframe between pivot points and trades the swing components of longer term trends. The intent is to trade price action momentum bursts (impulse moves in the direction of the long term trend or fast retracement moves of the long term trend). Once defined on W1, we then drill down into D1 to fine tune the schematics with more available data and fully encapsulate price action of the swing component using the SDC. This then allows us to define overbought and oversold zones relative to the mean of price action throughout the swing component (momentum burst).

The chart below describes the context of how I define swing components on W1 before I drill down into D1 for trade entry and fine tuning.



This strategy commenced in earnest in live walk forward mode on 18 April 2015 and will be added to the monthly performance summaries. The intent is to supplement the discretionary breakout systems with this retracement approach to supplement overall performance returns and flatten the equity curve.

So far it appears I am nailing the entries with a degree of sniper precision and we have yet to experience an ugly trade.......but we know they will come....so I am crossing fingers and toes for the future.

This approach is anticipated to produce a higher win rate than the breakout approaches described in this thread and also achieve some robust wins given the relatively high R:R ratio. The downside when compared to the breakout approaches is the lower trade frequency and I will obviously miss some great trades given the selective nature of the entry placement....still beggars can't be choosers. However as a supplement to the main game of trend breakout strategies, the approach will be a handy addition given it's relatively uncorrelated nature.

Each trade entry is divided in two with half reserved for a profit target of 3xR (if achieved) and the balance (or both halves) reserved for exiting on the trailing stop.

Below are screen shots of current live trades.











A few more to come next post.

C

New Discretionary Strategy Addition - DTTR (Diversified Trend Trading Retracement) cont'd......









That's it for current live trades.

Regards

C

Adrenalin rising....the natives are getting restless..........it's Friday night people's!!!!!!!!!!!!

Give me a beat............

What were the key issues discussed at the Hedge Fund Intelligence Summit 2016?

Not much..........

Enjoy a jocular blog post from Neils Kaastrup-Larsen who attended the conference and raised the most important question of the day to the Fund of Fund managers.

Trend Following Performance Report - March 2016

All good things come to an end......and what a cropper March was for the trend followers. Bring out the dead!!!!!!



It was just nasty. A significant component of the haul of positive returns since January evaporated with mean reversion shifts and general market uncertainty. The performance results of Fund managers in the trend following space all took a nose dive (with the exception of a few notables with composite systems where the day was saved from mean reversion strategies such as Byron Capital)......one valid reason why system diversification is becoming a more urgent priority for the FM's who occupy the trend following space.

Below is the status of the Trend Following Index as at March 2016. Despite the open slaughter for the month, the trend following FM's are still managing to keep their head above water for the YTD calendar result however the index is currently underperforming the S&P500 TR index for the period since Jan 2016.....come on trends.......where are you? Well don't despair because a little birdy is telling me that April is going to be a cracker.....however let's temporarily remain grief stricken and focus on March results.



So now looking at the equity curve above, lets focus on the March blip as we all know that trend following is dead.....right?.....now where did you say the blip was? That's right, when put into performance context, the March result is really nothing to get your knickers in a knot about. The TF Index is still significantly outperforming the buy and hold index of the S&P500 TR. So we just suck it in and continue to poke fun at the crusty passive buy and hold FM's who dominate the fund investment space. There is always a chance that they may be right after all. Maybe not in the last 14 years but who is counting? Investors live forever.

So now to the details for March........*pardon me* I am not sure how the music got in here but it must be gastric reflux from last night.



So here we go.



Have a quick squiz at March 2016 results. You will notice that of the total performance count of 50, only 4 Funds squeezed through with a marginal positive return for the month. Note that Conquest, Mark J Walsh, Quest Partners and RAM have not posted results yet for the month of March but I suspect that they are avoiding release to investors until more joy is received from April performance results....*cluck, cluck, cluck*. Anyway, I have given them a big fat zero return for the month which is a favour to them as it was likely that they achieved a miserable negative return for the month like nearly all of the others.

So onto the awards for the Month of March for the FM's. Will there be a major upset to last months best/worst performers?...........Well not really......I wish I could make this more interesting but the same old names just keep popping up. Here we go.

Adrenalin Junkie Award for the best CAGR between Jan 2002 to March 2016


Dreiss research (the full time surfer and part time FM) is continuing to knock the ball out of the park with a handy lead over Tactical Investment Management. While posting a great long term result, you need to appreciate that you will be in for a roller coaster ride if you want to enjoy these returns. If you want to know more about Bill's trend following post then the attachments to this post pretty well sums things up.......go the fractal.


Wooden Spoon Award for the Worst Drawdown between Jan 2002 to March 2016


This just makes my eyes bleed. Can someone please put this beast out of it's misery??? This is where discretionary trend trading can go wrong. Imagine the upbeat mood that would be experienced in this firm *sniggers* :-(


White Undie Award for the Lowest Drawdown between Jan 2002 to March 2016



This is more my style. For the nervous nellies out there (like me), this is the choice of choices. Pretty amazing result considering. Over a period of 16 years, the max drawdown was 10.59%.....just WOW!!!!!!


Bestest of the Best Overall Performer Award for the Highest Adjusted CAGR between Jan 2002 to March 2016


If you are after the best overall performer with respect to best performance return scaled to drawdown....then this is it...... The current investment god of the universe in the CTA space. Well done FORT.

Encouragement Award for the Best Annualised Sharpe Ratio between Jan 2002 to March 2016


Not my preferred performance category (as I prefer adjusted CAGR), but the Sharpe ratio is still a fairly robust metric that displays overall volatility of returns

Brief Moment of Glory Award for the Best Performance for the Month


Hold the horses......Byron pulls a positive return for the month of March. They must be on drugs!!!!

The Last But Not Least Award for the Best Performance over the last 12 rolling Months


Byron again!!!! That's just crazy talk. I want what they are having.


That's the wrap up for the month of March guys. Let's see the ship turn around in April.

Cheers C

You'r the Man.

You know I had to make it dramatic Bass....otherwise people just don't read the stuff

Swing for Swing Traders :-)

Aw....thanks mate... March has been an open wound for my automated systems....lest we forget. I will be sharing the performance of the automated systems in a few weeks.....but to bring a bit of joy, I will be posting the results of the Discretionary Systems for April which weren't too bad. :-)

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

Following the month of the living dead, April seems to have picked up on the trend front, at least with my discretionary breakout systems. I will be separately reporting next month on the recent addition to the Discretionary Systems, namely the DTTR discretionary system as this addition to the suite has it's own distinctive equity curve.

Anyway, let's jump straight into it.



Statistical Summary of Closed Positions


Top Trade for the Month


Current Open Positions



What's Trending at the Moment?

AUDCAD

Attached Image (click to enlarge)

EURJPY


USDJPY


Silver


To be continued......

C

ı have never see that much dicipilined trader.great performance.you are my role model

April Performance Summary - Discretionary Systems - Breakout Techniques (DTT, EDTT, EDTT1)....cont'd


DTT
An Ok result for the month with a performance lift that removed the small residual drawdown of March.




EDTT
A strong result for the month related to high activity for both in-trend and counter trend trades. A few more months like these and we may see the EDTT surpass the DTT in relative performance terms.




EDTT1
The poor cousin to the EDTT that trades multiple time frames but restricts activity to in-trend trades only. Still a pretty good month's result.




Relative Performance between systems
Despite the good showing of the EDTT for the month, in relative performance terms the simpler DTT is still shining. All 3 discretionary systems are rapidly accelerating away from the performance of of the TF Index and the S&P500 TR Index. While we are getting better at honing our automated systems, the discretionary methods still reign supreme.

Attached Image (click to enlarge)

That wraps it up for April.

Have a great weekend all.

C

Cheers V. Life is made a lot simpler when you automate as much as you can as you can pump out stuff pretty quickly. There are still a few things to improve to make performance reporting a bit less of a drudgery....but I am slowly getting there mate.

aveagoodweekend :-)

Hard work over................Penfolds Bin 8 in stock.......*glug, glug, glug*......it's time to party....so let's go back....way back :-)

Lol.. I can relate to that..

I'm starting to get the feeling you don't sleep... kinda reminds me of me..lol ..

Thanks for the Trend Following Performance Report, March 2016...quality analysis as usual mate and May already..wow.. where does time go..


I really like Bill Dreiss.. I find myself nodding in agreement in his interviews... I'm not sure if its the surfing thing which gives the persona he has a wild side or the fact that he wants to push the boundaries on exceptable performance ?... if you want 'on the edge returns' he's certainly an option, though strap yourself in.. Bills ride is a bit wild..

Just come back from a Japanese Cafe where I had the Chilli-Beef Shin...it was that hot I thought the top of my head was going to start spinning..I managed to save the day with copious amounts of vino though.. beats ham and cheese anyway..

That's just making me hungry. Copious quantities of vino tends to be a cure all. I can't recommend it more strongly. It does however occasionally make the speech difficult

PS Nice segue into ham and cheese BTW Ja :-)

Better Systems Trader - 048 Indicators, DSP & Cycles with John Ehlers

A tough going but interesting podcast from Andrew Swanscott that deals with the issues surrounding Indicator Lag and Market Noise - one of the biggest enemies of a trader. Some interesting viewpoints but their application to trading is not straightforward. Recognize that this discussion just scratches the surface of some interesting concepts.

John Ehlers - ElectricalEngineer specialising in electronic research who formed an interest in trading commodities, particularly specialising in digital signal processing and in the identification and methods of trading market cycles and in identifying and trading regime changes from cyclic mode to say trending mode. John started in fundamentals but quickly gravitated to technical aspects particularly focusing on the validity of indicators. John currently prefers swing trading and reversion to the mean due to the higher % of winning trades when compared to say trend following and in the ease of statistically testing these techniques.

Interesting Snippets:

1. A cycle in the most general sense is anything that repeats itself. In the context of markets, when you disturb a complex unstable system and you hit it with noise it causes the market to take off in a particular direction which generates a bit of feedback when it extends too far causing it to revert with an overshoot and so on and so forth like a pendulum for a temporary duration. Basically market data is very noisy and you can recreate it quite easily. Start with a random noise generator in say excel and then apply an exponential moving average to it (that EMA is caused noise with memory) and that is a reasonably good model of how the market behaves. There are cycles that come and go with various frequencies. They expand and contract....ebb and flow....but are a feature of disturbance as opposed to representing a causal effect over the course of the cycle itself......a bit like throwing a pebble into a calm pool of water. The initial disturbance is manifested by a cyclic wave pattern in the medium....a resonance from impact as opposed to a structured defined causal process. ;
2. MESA (Maximum Entropy Spectrum Analysis) - An algorithm initially pioneered and used for processing short term seismic data in oil exploration. MESA had application to the markets due to it's changing nature where cycles changed very rapidly...and hence you can only use a short amount of data. There is a 2 step process:
   1. Select say 10 bars of data (a length of data) and accurately fit a filter to that data (curve fitting to the extreme). You remove all cyclic information in that data with the curve fit solution and then you are left with maximum noise (entropy). Once this filter is created you take it offline.
   2. You then put a new signal source to the input of the generated filter to see which frequencies are generated by the filter itself..the spectral analysis..(generates cycles which are then cross compared to untreated data to identify the impact of the signal 'memory'). MESA is similar but a more effective method than say fourier transforms to interrogate cyclic and fractal market data as it takes into account spectral dilation that takes into account data scale;
   3. MESA can't be used directly but is useful in determining optimal lengths applied to filters. The intent is to determine the best length to use for an indicator eg. RSI/ Stochastic or other indicator. It interrogates and removes cyclical influences to determine the optimal period length of the indicator.

   
   

3. Cycles are frequently incorrectly used. People try to assign causality to cycles which is wrong. They then build a model to force fit the model to the data which is effectively curve fitting in the extreme. Assume cycles are there but they change all the time.
4. Market data is not a linear input and white noise is not additive to the inputs. Market data in John's opinion is 'noise with memory' or 'pink noise' that gets back to the fractal structure of market data. You do not need indicators to assess the periodicity. You just need to visually look at market structure to gain a best guess understanding of future direction. Do not get locked in to a given model. Don't make data fit the model.
5. If you use conventional indicators such as MACD and RSI etc. the usual treatment say on a typical 10 day trade is to wait for a signal and then further wait a few bars for confirmation. There might already be 3 bars of lag in the computation and then say 3 bars of lag in the wait for confirmation and then take the trade on the next bar. You are 7 bars into a 10 bar move......hopeless..it is a guaranteed loss if you wait for confirmation. The way to use indicators is to anticipate the turning point say 3 days in advance (to cancel 3 day lag of confirmation) and then you are right there on a typical 10 day trade and can get in at an optimal point.
6. Support and resistance lines are subjective and very difficult to program as are other measures such as fundamentals and seasonality. They should only be used by discretionary traders and NOT systematic traders due to complexities involved.
7. Any time you can increase the data set (introduce a new data stream) then this is good as it immediately reduces the signal to noise ratio.....except watch out for reading too much into indicator divergence.
8. Divergences/ Convergences - The trader must be careful that the added information is orthogonal...if it is just more price data you are not helping yourself.......WTF?...... (*Cop notes* I assume it means doesn't dilute signal strength through interference??)....let's break it down..... A divergence for example when tied to the same data stream as the indicator is a fundamental attribute related to the indicator itself and not the market data. It relates to the way you built your filter. An oscillator will not support a trend line for example. As the trend continues, the oscillator signal falls off (diverges) but it has nothing to do other than the calculation created by the filter. You need to use information from other sources (new data streams) such as volume or perhaps volatility to identify divergence.
9. Regime filters (eg from cycle mode to trend mode) - Big problems with them as if you use cycle analysis with the regime filters you are always in a wrong position when the market shifts into a trend. John avoids trading when a market feature such as a trend extends beyond his normal monthly cycle time horizon.
10. Hurtz coefficient or other auto-correlation indicators as a Regime Filter - These methods vary significantly in their calculation depending on how much data you use.....hence while being a useful concept which feeds into the fractal structure of the market is very problematic as a trading tool in themselves.
11. Smoothing Filters reduce lag so they should be used instead of simple moving averages or EMA's.
12. Favorite low lag filter and oscillator. Best low lag filter is called 'De-cycling filter' which cancels lag by taking the difference between two filters. Best oscillator is the 'Roofing Filter'.

Phew.....that was tough going.

C