Diversified Trend Trading Approach

Since the start of trying your approach I experienced profits (few percent a month), then drawdown from the start of the August....

How about drawdown recovery period now? Since all the scientific papers are praising trend following for relative fast drawdown recovery /tic

Hi C,

What about that situation, where hiccup would need to have more that 2.0 SD? Marked now on chart.... Would you go higher than 2.0 SD in SDC?

Morning R

This is where a backtest over a long range helps. The only guide you have available to provide expectations in regards to queries such as .....how long will it take to recover from a drawdown?...can only be inferred from historical performance statistics generated from your system....otherwise it is a bit like a ..how long is a piece of string question.

The best I have currently in relation to the techniques I deploy for the discretionary systems are as follows:

EDTT


DTTR


BOTH EDTT and DTTR


Over the short history covered in my walk forward tests using my range of instruments, timeframes and techniques deployed, the drawdowns have been up to 4 months duration....but the data set is too limited to form a firm conclusion from this. For example we have only really experienced a small sample of market conditions over this time.....and most of them have been the mean reverting environment post GFC. Just keep in the back of your mind that in the professional CTA space, most of the trend following FM's are approaching their highest drawdowns over a 17 year history at the moment....so it is clear that conditions more recently have not been favorable for trend followers. For example most of the CTA's have had a progressive drawdown build since May 2015 to current day..but what you need to watch is how quickly these drawdowns are building. If your drawdown is building in one month to say 20% or so...then this says that you need to put a cap on it and quickly and assess why it is growing so fast. You need to be able to survive for a significant extended period and preserve your available capital.



If you examine the CTA drawdown profile above then you quickly realise that drawdown is just a necessary evil for trend following.

What we do need to gauge however is what component of these drawdowns is due to symptoms of overtrading or error. These symptoms add to the severity of the drawdown.

Of course these equity curves are relative to your particular technique. Given that you trade different timeframes to me and instruments and also given the very limited time series we have to work on here....these are hardly sufficient to give us any ideas.

We will get a far better idea about questions such as these once we have performed our long range back-tests on our portfolio's and using our preferred timeframes.

What we need to remember is that the majority of our lives as trend traders....we are in drawdown. That is just a feature of trend following. The times we reach blue sky are only during those rare unpredictable times when momentum shines. It is during these periods that we hit new heady heights and pay for the past non-trending periods.

Given that we are predating on rare events our mindset must be one of capital preservation at all times which means that it is more prudent to wait for a trend than to assume that every trend will achieve our objectives. Our win rate is about 40% and about 90% of these winning trades are only marginally above breakeven. It is only the rare exceptions of 10% on 40% winning trades that make life worth living.

What we need to be particularly careful of in deciphering our drawdowns is being able to distinguish between the drawdown component resulting from market conditions versus the drawdown component associated with over-trading and errors. These symptoms can only be detected when comparing against a benchmark such as a historical back-test that has stood the test of time. The longevity of the backtest across a broad range of market conditions tells us that the system deployed is robust in terms of position sizing and risk management and that the parameters have been set appropriately. If we therefore follow the trade plan associated with this historical system to the letter we can ensure that our actual drawdowns going forward do not possess a significant component arising from overtrading or errors.

It is essential we have these tests in our analytical arsenal for without them we are like a rudderless ship.....and will get no closer to being able to make an informed answer to our queries such as these.

With DTTR you are seeking to define the longest range trend you can on the timeframe concerned, as what we are seeking with our logic is to define ideal entry points on retracement that offer excellent tight position sizing and thereby enjoy great returns if we get a nice retracement back into the longer term trend.

I don't have H2 but I have used H4 on AUDUSD to display what I mean in regards to your query (refer to the chart below). I think that this swing low point is also valid for your H2 chart.

I was able to define a swing long way back at 2017.05.09 which creates the longest look back I can use for this timeframe to provide an indication of the bias available in the time series.

When I draw my DTTR SDC's it is visually evident to me that we have not yet had a confirmed *obvious* trend change from the prior long bias. I would therefore prefer to have a long conviction as opposed to a short conviction with what the historical data series is telling me. I have therefore slightly widened my SDC's to incorporate the new most recent data and would have been in a position to take a new entry long at 16.08.2017 as conditions for entry had been reached.



I hope this makes sense mate. The DTTR is a long range system that attempts to define bias in a more quantifiable way than the breakout techniques and defines overbought and oversold zones in the context of the historical data series. No matter what timeframe you use, ensure you use the maximum available data to have a chance at detecting what is bias as opposed to what is random noise.

PS With regards to the DTTR a conundrum is faced whenever you are at a possible trend turning point. Whether to go long or whether to go short. I try each direction when the opportunity presents itself....but only once each otherwise it is best to have a breather and wait for more data. You could find yourself in a yo-yo bleeding yourself dry if you keep trying to second guess the direction around these points. Obviously the trend turning points are attractive as they offer potentially great entries for some wicked long term runs with great position sizing....but there is a cost in focusing on these which is a potentially low win rate.

DTTR recap - The logic underpinning the strategy

What this technique seeks to do is identify ideal entry points into very long term trends that offer great risk:return ratio's if they work out.

For example below is a chart on W1 for AUDUSD



What we want is a strategy that identifies optimal trade entries into these nice long range trends.

Obviously hindsight is a beautiful thing, but there are places at some retracement locations on these big moves I would have loved to have had a position on with tight stop losses and excellent position sizing.

Here is how my hindsight plots these various locations. I would be sitting pretty with humongous profits if hindsight was my guide.



So now to translate a strategy that has a chance of getting close to picking these entry points.

I won't go through the entire exercise as it is a time consuming process but below is how I catch the first entry point as a guide to how we apply the SDC's to achieve this.



You need to practice this going from left to right as unfortunately hindsight is not available on the right hand side of the chart.....but hopefully you see how the logic is being applied. We need the longest available trend we can draw to determine overall directional bias in the long-term, then we want razor sharp entries at classic reversals with tight position sizing. You will have many failures risking minimal capital each time....but when you nail them.......you are king of the world :-)

Thank you for the quality explanations, as always.

How do you see such EUR JPY short

Yes, I need to practice more on the simulator....

I fear not the man who has practiced 10,000 kicks once, but I fear the man who has practiced one kick 10,000 times.
Bruce Lee

100 x 30 minutes bars correlations..... If you trade CAD JPY, you are trading indexes /tic

I like the logic mate. That works for me. Just remember that now I am in defensive mode I have pushed pending entries further away. I also stagger entries between the timeframes a bit more than you. I ensure I am deep in profit and above a worst case breakeven before I enter in the same direction on another timeframe.

Here are my charts for EURJPY on EDTT and the logic as it stands now.



On MN the primary trend is long and has been since April 2016. Price appears to be mid-range on the MN SDC. It could either go up or down in the shorter timeframe.

On W1....same deal as MN.

On D1 it looks like the long swing move has come to an end and we might be going short.

Based on the 3 timeframes above, my preferred bias has shifted to a short viewpoint in the short term for H4, H1 and M30.

H4 is recently short but needs a greater data set to get me excited and it needs to be staggered after M30 and H1 to induce me to go short.

H1 I have extended the SDC to include more data from the swing high on 2017.08.02. This is looking good for a short entry into a trend of moderate momentum.

M30 My swing high on M30 is at 2017.08.06. This swing has higher momentum than H1 and we have less data to run off. I would like more data before I commit to jumping on board this faster mover.

Now I make a choice between H4, H1 and M30 as to which I choose for a short entry. Of the 3 candidates H1 is my best bet IMO given the lower momentum required and the increased data set. My next preference is M30 so I have staggered the entry short to coincide with me being deep in profit on H1....and last but not least H4 as we need more data for confirmation.

I hope this helps R :-)

EDTT - Preliminary Results of Medium Range Tests - Discretionary Systems

I have been spending the week developing some analytical tools to use with a great back-testing kit that was referred to me by R. The software allows me to quickly and accurately conduct back-tests within the MT4 environment and cross-compare results against my live walk forward tests I have been conducting since Jan 2013.

While it is still early days in that of the total universe of instruments (24) traded under the discretionary systems in this thread, only 2 medium range back-tests have been completed, the preliminary results auger well regarding the sustainability of these approaches.

The medium range backtests commencing from 1 Jan 2010 to current day provide an enhanced data series to assess important performance metrics such as optimal position sizing arising from my maximum drawdown tolerance.

The results of each instrument can be assessed in isolation or as part of a portfolio.

In going forward in this thread I will be reverting to using statistics derived from this back-testing suite as opposed to maintaining the live walk forward tests which have been periodically reported in this thread. This will save me significant time in report preparation. I will still be monitoring the programs on a live walk-forward basis to cross-compare against my backtest results to keep sanity in check..however due to the time savings....I will not be regularly reporting on the live walk forward programs.

Anyway....without much further ado....the results so far on 2 instruments tested since 1 Jan 2010.



Note: The following results have been scaled within my maximum risk tolerance of approximately 20% during a very difficult trend following period.

EURUSD

Annual Results


Summary



EURJPY
Annual Results


Summary



Portfolio Blend - Currently only 2 instruments
Annual Results


Summary




Comments
While the performance results are strong, clearly the volatility of returns need to be dealt with. As more and more instruments are added to the portfolio we will see how we can smooth things off to create a wicked portfolio performer.

Have a chilled weekend.

Stay posted.

Cheers

C

Who would have thought, that from my small remark about simulator I use, such great thing will be created by C.... sign of bright things to come....

I owe you a few beers R. I am happy as a pig in mud at the moment. The simulator is going to take a huge chunk off my workload :-)

Something special from Australia ....

Yes and it allows to test 10 years of data without waiting 10 years. Time is the most precious commodity.

Phenomenal BN.....I never knew she existed. She gives Mike Oldfield a run for his money in instrumental/digital diversity. A solitary jam session using the power of loops :-)

How's things on your side of the fence mate?

It’s good to see his video when he was young, It’s amazing to see how body transform in time yet voice remains the same, all different path taken will get us to top of the mountain, Bless are those who embrace the philosophies of the past and welcome the Digital age philosophies and lucky are those who can see what yet to come.

Alan Watts - Full Unedited Version of 'Conversation with Myself'

Hello, im sorry to post here but i could not stop myself from posting this result of EA trading, started the account with 5250$ on 25.5.2017 and final result as on 18.8.2017 is 90000$ in profit. withdrawal of 69000$ also attaching the report of trading.

Nice stuff Carp.

It definitely is an Alan Watts chill-out Sunday :-)