Diversified Trend Trading Approach

THE RECAP continued....


Commence with a clean chart in bar chart mode and pick an instrument that you would like to include in your portfolio. This technique applies to any of the following asset classes namely forex, indices (CFD's) or spot commodities. I can't say whether it works on equities and other asset classes simply as I haven't tested this technique on them but...... it appears evident that this technique is 'instrument agnostic' which is what you want with a robust durable trend trading strategy.

We will choose for this recap good ol' USDJPY....but in reality any would suffice but..... I have chosen this (in hindsight) as I know what the following recap will demonstrate and for the purposes of learning this technique we want to know how to trade a trend.....and we have a cracker in store for us. But remember....life is like a box of chocolates......you never know what you're gonna' get.

Just remember that this discussion will focus only on one instrument......keep in the back of your mind as we step through this that in a diversified portfolio you could be looking at over 20+ instruments. So speed of execution and decision making is going to be essential and for that, the correct platform for this technique 'really matters'.

Also remember this as we step through the recap. This is not an elegant trend following technique with exact lines and formal patterns that would be traditionally regarded as a trend. Such an approach is counter-productive in trading a diversified portfolio as you need a technique that is fast and identifies upwards or downwards price movement of significance...no matter what the pattern. This technique is one of brute force that works well in stochastic markets and cuts through any illusory pattern you may think exists, to deliver absolute returns by only focussing on the barest rudimentary pattern of all....and that is whether price is going up or down in a concerted and sustained effort that may deliver a 'real' trend of substance.

So a trend in this approach takes on a different form to that which traditional trend chartists may agree on. This is a no b...hit fast punch-through technique where considerable discretion is allowed for. We are not dealing with exactness ladies and gentlemen as this market exhibits significant random behaviour but what we are hoping for is that in this general random noise there perhaps is a 'sniff' of a kernel of co-ordinated price movement deep within the price pattern.

So don't worry about exact placements of channels and predictive entry points as these allow for considerable discretion. In managing a diversified portfolio, it pays dividends to not be too anal about things. All we want to ensure is that we participate in any significant price movement and therefore we are not too selective in how this directional price unfolds. As a trend trader who depends on directional price movement, we want to enjoy as many as practically possible as in the scheme of things this market tracks more sideways than it does in the vertical dimension. You just want to avoid where possible that damned sideways movement, or what I refer to as price congestion.

The recommended template for a proposed 'single' instrument is below.



Note the following:
1. Clean chart with only information made available that is helpful to this technique. Forget any other additions. It might make a less elegant chart and give the appearance of 'too simple to be practical'...but in essence that's just what we want. To be brutally honest, we could also be working just with line charts as opposed to bar charts as that really is all you need with this brutal technique.....but I personally prefer bar charts. Note that there are no candlestick charts. Their use is not a problem but their colour and the unnecessary information, that you won't use anyway for this technique, quickly gets these charts unnecessarily complicated and visually dificult to interpret. Speed in application is an absolute necessity in this technique, so please make it simple for yourselves.
2. The timeframes we monitor are monthly (MN), weekly (W1), daily (D1), 4-hourly (H4), hourly (H1) and 30 minute (M30). You will need to develop a template for each instrument that allows you to quickly switch between them wihout recreation.....and if you have a portfolio watch-list like mine....there will be 24 templates...plus a summary template (which I will get to later). "Oh my god (I hear the cries)........that sounds too hard"......na, it's not, and familiarity over time makes management easier. One click and a new template is up. That's how it should be ladies and gents.
3. "What is that platform you are using"......ProRealTime
4. Why not MT4....... Personal preference but also because my preferred platform offers 'trade by chart' functionality and a very user-friendly interface...but this shouldn't stop you using MT4 if you prefer. This technique works on any platform offerring 'Standard Deviation Channels' but you need to ensure it 'works for you'. I find MT4 not as user friendly as ProRealTime, but that's just me....and I am a teapot anyway!!!!/ tic

So now have a look back at the USDJPY template above. Before you do anything further, make an assessment visually of what timeframes are trending. I have synchronised all timeframes to be representative of a point in time in late June 2014. Get the feel of overall price movement.

What we are going to do now is ensure we are all defining that price movement the same way, not by personal chart reading but by applying the only indicator you will use in this technique...being the Standard Deviation Channel (SDC). By using this device, we ensure we are all 'see' the same thing with price movement. This avoids any subjective judgement that may be deceiving you.

Next post..applying the SDC to reflect a possible order that exists in this chaos that we can all agree on.

C

THE RECAP continued....


Application of the SDC on all timeframes to determine..."what's goin' on".


Error to chart above. The colour scheme on the M30 timeframe should be beige....not purple, which is the colour scheme I adopt for the H4 timeframe. Sorry about that chief :-)

Ok now this requires some thought while recognising it is not an exact science.

I simply choose high and low points on each timeframe that represent the commencement and end points of general price moves in the upward or downward motion. Now even using this principle we still run into interpretational problems as the chart reader who is pedantic will be insisting that a price move actual commences *here* and ends *there*. Don't worry about this precision but heed the following advice......the longer you can determine general directional price movement, the more valid is the information contained therein of the general nature of any possible 'hidden' directional order in that chaos. So where you can, stretch it out. If you have two low bars that are seperated in time and differ in height with the most recent actually being slightly lower or higher than the preceding bar, then use the preceding bar. What is lost in exactness is made up for in creating a wider SDC which gives more room to breathe and is a necessary pre-requisite in dealing with unpredictable price movements.

HINT: Is this hopeless search for exactness a fundamental trap that all heavily rule based systems and EA's fall into when in fact in this highly random noise of the market, no such exactness ever exists?....now that's something to mull over. So in the absence of exactness, isn't it best to use the tools of probability itself to determine the most 'likely' action to take. I will leave you all to ponder that one.

So refer back to the chart with fresh ideas and look at where I have extended my SDC's from and to. Note that I am extending from a broad interpretation of a high/low to a high/low in the information that is made available to me. I will stretch out the chart deeper into the past if necessary to ensure I am dealing with a single phase of directional price movement hat can make some weird patterns in unfolding. They key thing is that the SDC will create order to this weirdness.

Q&A:
1. What settings are used for the SDC on each timeframe?........A setting of 1.0 *thanks genie* width with each SDC drawn having a distinct colour that allows you quickly identify the timeframe (sorry to those with colour-blindness out there...but choose your own distinct visual identifiers of a distrinct timeframe).
2. Why a setting of 1.0?.........The width of the SDC signifies the probabilty of closing price on a particular timeframe to lie within a statistical variance from the mean of that price range by a certain percentage. With a width of 1.0, the data distribution (of closes) is (with 95% confidence) representative of 68% of the data values. A width of 2.0 represents 95% of the data values....and 3.0 represents 99.7% of all data values.
Sure to be safer than sorry we would love to have as wide a SDC as possible to give room for future price movement to breathe but with this technique, this of course correspondingly reduces trade risk% (which we will get to) resulting in smaller position sizes. A width of 1.0 is in a sense optimised for this method of trading where you want to make hay when the trend shines but avoid leaving too much profit on the table for the flock to enjoy....and of course this means that more frequently you will be whip-sawed prematurely out of trends but you can always re-enter them later and some of the longer term timeframe trades will just sail through it as their 'relative' SDC is wider in the sense that the range between entry and stop is considerably larger even at the same width setting of 1.0...but also remember you need to 'own this strategy' so choose the setting of your choice after you have strenuously tested it empirically.
The width of the SDC is defined by price volatility around the mean and this provides an excellent indicator of average true range. Accordingly the width of the SDC is used to define your stops so in volatile markets, the width of the SDC is wide for 1.0, and in less volatile markets, the width of the SDC at 1.0 narrows. You will see that over time in it's application the SDC 'breathes' to the tune of the market just like a Bollinger band....and that is not surprising considering the most distal end of the SDC actually represents a moment in time of a Bollinger Band.
3. How is the SDC different to the Bollinger band.....The SDC only concerns itself with price movement within the defined extent of the channel and as we draw the SDC from high to low or vice versa, it therefore is only concerned with the price data of the trend itself. It excludes those periods before and after the trend is identified unlike a Bollinger Band that simply looks at X bars back to define the statistical variation of price irrespective of whether that price series is trending or not. In riding the trend, the SDC gives me greater relevant info relating to trade management (which we will progressively see in this recap) than that delivered by a Bollinger band.....though I still really really like them anyway.
4. What is the centre-line in between the outer extemity lines?......That is a very significant line which is the mean regression line (or linear line of best fit) associated with the price history contained within the entire channel. This is the line that represents whether price on average is going up or down and it's speed of movement (represented by the slope). A steep line is a fast trend usually of shorter duration than a trend with a lower slope....but remember we want to be riding a trend irrespective of thr slope and some of the better trends to ride are those with a progresive build in momentum over time. They may start out as trends on original low slope that progressively pick up steam over time. I actually prefer a less ambitious slope than a fast trending slope as the former tend to have longer duration and ultimate directional movement in price extent.....but you really can't afford to be picky. Up is up and down is down and that's the way you need to treat it.

In a nutshell guys, the reason why I don't use a traditional trend following technique (even though there are some crackers out there...thinking Fartist, Graviton, Stevepatt...) is that I am after a tool that deals with probabilities and not exactness. Namely the Standard deviation Channel. This makes me enter more times than other techniques but with miserely risk management we can deal with that and ensure that as soon as a trend sniffs, I am onto it. :-)

So now, look back at the template above. What is it telling us about the recent history of price movement. Which timeframe is trending and what is it's speed (or momentum) and volatility ?


Firstly you should be looking at the highest timeframe chart to see the direction of the primary trend. We will only be trading in that direction. Quite evidently the primary trend direction in this example is long (or up) so that's the direction we will trade on all timeframes.

Q. What if the MN timeframe isn't trending and we can't define a direction of the primary trend?.........A) then go to the next timeframe down to determine the primary trend direction and so on and so forth.

So, back to the example...we have a primary trend direction which is up....good, so we will be going long on all timeframes.

The W1 timeframe is also up......that's great. It is also quite convincing and appears to have good momentum.

The D1 timeframe is a mess!!!! What's up?....well in attempting to draw your SDC's, there is a lot of confusion in where to draw them from. It's up and down and up and down.....*aaaarg!!!*. That's a sign that the instrument on that timeframe is range bound. There is very little price information you can extract that gives confidence that a valid long trend also exists on this timeframe in recent history. What I therefore do is extend the SDC across the entire section (hashed lines) to let the SDC attempt to work out general price direction across the entire series. In this example, there appears to be a small short (down) bias in the price series....so nothing to see here at the moment. Avoid this timeframe until solid evidence for directional price movement 'long' emerges.

The H4 timeframe is short (down) but it's momentum in reality does not appear to have much oooomph. Also this is counter to the primary trend direction....so avoid until the direction changes.

The H1 timeframe is also down but appears to have good momentum...however...yuck....avoid as it is counter to the rule of this technique.

The M30 timeframe is down as well and boy that looks fast. Should I be shorting?........No, because you will then be counter-trend trading and it is likely that what you are seeing is a simple retracement move before the next leg up on the higher timeframes....but yes, you might also be lucky and be at a trend turning point on the higher timeframes and therefore possibly entering at a nice early position in the future downtrend....but that's predicting and not in accordance with the rules of this approach so stop that brain doubting the technique and stick to the strategy.

Putting D1, H4, H1 and M30 together, these are representative trend retracements of the primary trend represented by MN and W1. Now look at MN and W1 and see which part of the chart represents the price history represented in the shorter timeframes.

Ok, so we have our SDC's drawn and we now have a feel for how the market is trending on different timeframes. Now it's time to get cracking and work out my pedictive entry placements to ride these suckers.......but lunch calls. I will be back to continue........

Any questions so far for my return? Just nominate the Post number and the query that relates to it.

Copernicus, you decided to change to a 0.5 SDC instead of 1.0? Seems so sudden. Care to go further in depth behind your reasoning?

THE RECAP continued....

I'm back....and boy is Brisbane turning on the weather today. It is raining cat's and dogs.......literally. Pretty rare event in March as traditionally this is beyond the typical 'wet season' we get here. The number of cyclones late in the season we are experiencing in QLD is an omen....but the old timers would be saying....it used to be this way....and I just put that down to a bit too much aluminium. :-)

Ok. So now onto..........

SETTING PREDICTIVE ENTRY LINES
"Hold on a minute Copernicus....you said this wasn't about predicting? Just ride the trend you say".........Ok Mr wise guy, in reality what these lines represent are simply reference points to indicate that if price reaches that level, to redraw the SDC and see if a valid trend appears to exist. I therefore give sufficient space on the relevant timeframe to allow a suitable price mattern to emerge so I can shout 'trend ahoy' and if valid then jump on board. I am a surfer out in the market sea and I am positioning myself into a location where if a wave surge exists, then it should be substantial enough at that place for me to ride it. I actually have no prediction or idea if price is actually going to reach these lines at all. They may or may not...but I am visually and figuratively positioning myself to take advantage of them if they do.


Note that the chart above has moved forward in time to the chart presented in the prior post. You may be wondering how the hell the M30 trend slope is now up and that is because the charts above have moved forward in time.

So what goes into my thinking in determing that place to sit in wait?

Firstly we are in-trend trading so on each timeframe there is only one predictive entry line. Not two as in the DTT or EDTT. That predictive entry is positioned for a break-out trade where a new high or new low is clearly reached in the relevent timeframe and in accordance with it's position ahead of the current SDC line you have drawn on your timeframe. For this example, the primary trend is long so all the predictives are place for a long trade only.

Why don't you look at retracement entries. You get a better price on entry?.......well the main reason is that being selective on entry like this means you may actually miss some moves if you are not able to jump on board with such techniques. Once again this is a brute force entry with minimal decision making to ensure we can get on board every trend of substance. The point is that in this technique 'a sniper entry' is not needed but most importantly you don't want to miss the wave as it could be a while before they return.

I frequently set pending orders to buy at this point if I am going to be away from the screen. However it's nice to at least confirm the nature of the historic price movement immediately prior to this preferred entry point so if it is clearly haphazard in nature with extreme volatility on say a single bar or merely a few.....you want to avoid these one's. In these instances, then simply put your predictive entry further away to allow time for the volatility to settle down or obtain further price data to more exactly define this volatility with greater precision.

"Are there any rules here?"....well not really as more importantly it is experience that counts in that respect here. Get familar with the SDC and entry placement and you will start to get a feel for it. Ideally you are after consistent prior past action with directional bias and lower volatility....but once again....don't get to anal with this. Just avoid the really obvious volatile nightmares that may comprise a single 'finger of death' price movement and are not truly representative of the entire time series of price action that has occurred in the channel up to this point.

Note also that these predictive entry lines give a bit of room above previous highs and lows....well that is because I am suss about false breakouts and it is a legacy from direct experience I have had to deal with. I am prepared to give some profit away than get too lean and mean with my breakout entry.

So now look at the template above and the overall scheme of things with predictive entries.

Note that the M30 predictive entry is close to where current price action sits. If price reaches this point to trigger a buy, this will be he first position to be activated......then a bit later the H1 entry.....then the H4 entry......then the D1 entry.....then the W1 entry....then the MN entry.

This staggerred entry across timeframes is going to have a similar, but better effect than traditional pyramiding on a single timeframe. The reason for this is that the defined stops on each associated timeframe are configured for the trade risk% asociated with positions taken on that timeframe. As a result, and this is important....you may get knocked out by temporary price whip-sawing of your positions on the lower timeframes while those positions on higher timeframes are preserved. All that means is that if trend resumes in the primary direction on the lower timeframes, you simply climb on board again while at the same time enjoying the additional profit retained on the longer timeframes. The traditional pyramiding technique on a single timeframe means that when the last position is knocked out, you close all positions remaining on that single timeframe to lock on a defined (but possible lower) profit...and then if trend resumes you have to start all over again, building the pyramid from scratch.

Now we have the predictives in place and the SDC's established on all timeframes, we are ready to rock and roll and see what this market delivers.

But note here ladies and gentlemen, that we are starting from scratch at this point in time. We have come to the screen with no existing positions in place and we are starting this technique from ground zero. Here is what to expect. Whenever you shut down the portfolio or come in from scratch to use this technique, expect losses early as it takes time to get into the rhythm of the market. Unfortunately it is a fact of life. That's why when I shut down the portfolio in January and returned shortly after to re-establish it again, the first few weeks were 'painful' as a number of losses were incurred particularly on the short timeframes before I got into the rhythm. Also the SDC from a prior loss is often preserved if price returns in trend after a whip-saw quickly which allows you to extend the SDC further to include more price action and better validate the directional and volatility nature of the extended price move.


Ok. I am going to leave things here for a bit. The recap is going to be a longer exercise than I anticipated but if you can be patient, it will ensure I can attempt to anticipate any possible queries and incorporate them into my discussion. It may take a few weeks and I would like to restrict this recap to my Saturdays. It is quite exhausting and furthermore it gives everyone time to digest this more slowly and bring up questions and queries before we progress.

Farewell for now. I apologise for this delay but I will temporarily be returning as a git on this forum generally and will re-commence next Saturday. I will also be re-reading my prior posts and further amending and editing if required......and also responding to queries if required on the recap so far.

The post next Saturday will commence where we have just left off. Where we will start trading this sucker if probabilistic price action confirms entry. That's where the real fun begins. Stay tuned.

Take care

C

Hi genie

Sorry, your post slipped through the cracks but thanks for the question. :-)

EDIT EDIT EDIT.....Christ genie you are right. It is 1.0 SDC initially on entry. I will amend all prior entries in the recap.....thank God for observant eyes. Don't know what was in that dense head of mine. I took your question to actually be double what was intended but that was actually 1.0 SDC.....How embarrassment. We do actually adjust when we are deep on profit to 0.5 SDC but that will come later in the recap.

genie can you please just cross-check that I have now so far removed any reference to 0.5 SDC in the recap? I think I have amended all the errors but better to be safe than sorry. Much obliged......





C

Weekly Outlook

Ok guys here is the regular weekly outlook. Just remember however that this is not to be considered part of the recap. The technique applied here relates to the EDTT which trades both in-trend and counter-trend to the primary trend.

Given that we should now be getting the hang of this approach, I won't be including a written assessment of the outlook ahead. That is now for you to interpret yourselves based on the charts presented. It saves me significant time in presenting them this way.

I have already expressed the activities of the week in prior posts and in particular the impact if the FOMC announcement on the portfolio so no need to give an update on the week that's been.

So before we get into it....the necessary music to provide a background to interpret the charts ahead. An oldie but a goldie.





ASX


AUDCAD


AUDJPY

Attached Image (click to enlarge)

AUDUSD

Attached Image (click to enlarge)

CADJPY

Attached Image (click to enlarge)

DAX

Attached Image (click to enlarge)

...more to come.

C

Weekly Outlook continued......

DJI

Attached Image (click to enlarge)

EURAUD

Attached Image (click to enlarge)

EURCAD

Attached Image (click to enlarge)

EURJPY

Attached Image (click to enlarge)

EURUSD

Attached Image (click to enlarge)

FTSE

Attached Image (click to enlarge)

....more to come.

C

Weekly Outlook continued....

GBPSGD

Attached Image (click to enlarge)

GBPUSD

Attached Image (click to enlarge)

Nikkei

Attached Image (click to enlarge)

NZDCAD

Attached Image (click to enlarge)

NZDJPY

Attached Image (click to enlarge)

NZDUSD

Attached Image (click to enlarge)

....more to come.

C

Weekly Outlook continued....


USDCAD

Attached Image (click to enlarge)

USDCHF

Attached Image (click to enlarge)

USDJPY

Attached Image (click to enlarge)

USDSGD

Attached Image (click to enlarge)

XAGUSD (Spot Silver)

Attached Image (click to enlarge)

XAUUSD (Spot Gold)

Attached Image (click to enlarge)

That's it for the week ahead guys. Some trends are still active following the FOMC so it is not over till the fat lady sings and some new opportunities are opening up.

Trade well.

All the best,

C

Yup, all good

Regarding post #361, Point 2. I know you're not using MT4, so just wondering how your charts and different timeframes for each currency pair is currently set up? Obviously having 6 timeframes open for each pair would result in having 100+ open charts. Just want to know how you have them set up, and if I can transfer that logic to MT4. I would like to have predictives in sync within all timeframes of each pair, but dont know if possible on MT4. Could be a simple solution that im overlooking. Thanks

Thanks G :-)

Now comes the dillema...*thinks*. When I was using MT4 on an earlier derivative of the DTT, I was ok with using up to say 10 charts using a template on a single timeframe. I never really progressed past that with MT4 as the capabilities of ProRealTime made the decision an easy switch given it's user-friendly nature with this technique.

How ProRealTime works is that it is utilises templates each comprising as many charts as you have screens for but you get up to about 28 or so templates that are made available. So while I may need to work with say 100+ charts in total, each template once activated restricts the charts open on the screens at any one time.

Another benefit of the template system is that ProrealTime progressively includes information of higher timeframes onto lower timeframes so it always synchronises the on-screen info working across multiple timeframes. That means that once you get familiar with this technique you can have a summary template of the 24 instruments on the lowest timeframe and actually see the predictive entries and inclined trailing stop lines on each instrument of all timeframes. As a result, most of my time is spent on this consolidated template just keeping watch over the portfolio.A sample is below of the consolidated template which on a good high definition screen is easily interpreted and managed.



When an alert or visual signal is activated in ProRealTime, I then immediately switch to the relevant 6 chart template (using a drop-down list of all available templates) and find the instrument concerned to drill down into the relevant timeframe.

Given that MT4 is incredibly flexible, an advanced user with skills in this area would be able to give you better advice on whether this can be achieved.

...if the verdict is no....then there is a decision is to be made. Do you switch to another more suitable platform if you want to trade this approach? For me and my confidence in this system, the verdict reached is easy...but to convince others....now that's the question.

I sometimes wonder how attached we get to a particular platform at the expense of performance....t'is a good thing to mull over. Fit the platform to the system and not the other way round otherwise performance is compromised. The question I need to find out is what do the best performing diversified trend following fund managers use as their preferred platform and go from there?

I hope this doesn't put you off. You might be a very bright spark that can deliver this outcome on MT4....at least with help from others which I hope is possible.

I am not sure whether this answer helps or not G but at least it poses some thoughts to consider.

PS The trade by charts functionality is also another amazing feature of ProRealTime and if you are a TA enthusiast, it packs a whopping punch. However for the EA developers amongst us...MT4 appears to be the way to go. In this matter, EA's are not really relevant so far for this technique, as there are major obstacles to overcome in developing rules based systems around when the SDC begins and ends.

Regards

C

PPS Genie..there are other factors as well to consider that may influence the platform you use with this approach. For this approach to really work...the key pre-requisite is UN-CORRELATED DIVERSIFICATION (or at least reduced correlation) across asset classes. What this does is therefore challenge the offerring of the broker you use. My broker (IG Markets) has a large offerring but still is deficient in spot commodity assets but hopefully with enough whinging from clients...this may improve. They have a huge offerring in commodity futures...but that is a bit more problematic for me given my approach....but I am working on this. Given that my broker has 'sufficient' instuments to trade to give me at least a degree of un-correlated diversification PLUS ProRealTime as a choice of platform, the decision is easy to me. You may have a broker that offers this broad range of instruments across asset-classes but if they are simply FOREX specific, you may not have as much joy with this approach as I do...unless of course you trade using multiple brokers which may be an option for you.

Forex instruments as an asset class are fairly highly correlated so that can influence the risk profile of the approach. I therefore recommend spreading the joy across asset classes to reduce this possible risk.

Copernicus, this is doing my head in. I was up most of the night thinking lol.

I went through some many options/brokers and their platforms etc.

What would be ideal, is what you said, a broker that offers a diverse range of assets and has a very good platform. It is hard to find a broker that offers everything, such as sugar, wheat, etc, but Oanda is one of them (though they use MT4). However, I dislike trading with a MM so that option is out.

I know how important it is to trade a diverse range of products for this strategy. Really lost here and dont know how to proceed... but thanks for your help so far.

Genie.....don't lose heart. Where there is a will there is a way.

For the purposes of becoming familiar with the technique keep using MT4 if you like as that is a fine platform to hone the skills and develop your confidence. For example you could cut your teeth using say the DDT approach alone or better still this enhanced EDTT1 approach (as detailed in the recap) using a reduced portfolio size of even single timeframe.

My instinct is telling me that this approach being 'instrument agnostic' is primed for use in futures trading and this is where my own thoughts are leading me as well.

In the world of futures, un-correlated instrument diversity, leverage, depth of instruments and platform availability is a standard and a different kettle of fish (and that's why the CTA's love it) and with a bit of effort from me, probably the ultimate world to apply this technique. I just need to study it a bit more and also obtain a program that stitches a future's instrument's time series together to allow me to test and apply this technique. I know this exists as the book "Following the Trend: Diversified Managed Futures Trading, by Andreas F. Clenow" describes this possibility.

.....anyway I hope you keep interested and don't get frustrated with the perceived obstacles in the way.....unless regulation forbids it, there is always a legitimate way around these things. :-)

C

PS An email or phone call to Oanda might find that they do allow 3rd party platforms to receive their data. If that's the case, then you have a solution as ProRealTime is a platform developer and through direct contact with them, you should be able to obtain the platform and if possible, connect to a feed from Oanda. It's amazing what brokers allow when they can snare a client who is an active trader *sniggers*. Furthermore, I haven't checked out the capability of other platforms like Ninja Trader etc...so it is worthwhile just an an interesting exercise in itself to check these out.

We know the DTT approach so far is a cracker and you can get by with an excellent result from a far narrower range of instruments and timeframes. I am hoping that the EDTT1 is a 'super-charged' variant as I am a greedy b**gger, but time will tell with the tests. You may find that you need to commence with a 'Ford' before you can upscale to the 'Bugatti Veyron'.....but heh....there is nothing wrong with a 'Ford.....except......."


PPS There is always Graviton's, Bassramy's, Fartist's, Steve Patt's etc. techniques to consider as well and their superb trend following techniques are honed for MT4 but *sobs* I would be sad to lose you.

Thanks C... I'll figure it out

No probs G. :-)

So now onto.......... SETTING PREDICTIVE ENTRY LINES

Hi Copernicus
I have been trying to work through your re-cap approach on the EURUSD during 2012 and 2013 using MN, W1, D1 and H4 time frames. Is it possible to identify the trade time frame of the back test trades in the spread sheets? I know the back test doesn't include 2012 but for 2013 it would help to see where your entries and exits were in relation to the SDCs I am drawing.

Also, on the re-cap chart for USDJPY H4 why have you selected the level of the predictive line at 103.51?

Hi Koglkama :-)

Back in 2013 we were using the DDT approach and the trade transactions for that year are detailed in that big spreadsheet found in this post. It's safe to download so just unzip and look for the tab which includes all trade details.

The entry price will reflect the predictive entry line on the H1 timeframe. Remember that back then we were trading both ways (intrend and countertrend) on a single timeframe using D1 as a guide before drilling down into H1 for a entry. I hope this helps. Unfortunately I have not kept all charts so you will only get one of the predictive entry lines that are represented by the entry price itself.

With regards to your second query relating to this post, I ensure that each entry from M30, H1 and H4 etc. is staggerred so I am not hit with multiple entries at the same time (for trade management purposes). As a result, condition 1 is that the H4 predictive entry needs to slightly higher than the lower timeframe (H1) entry.

The following process I tend to also do is really unnecessary (and I recently disregarded the use of support/resistance for example) ...however bad habits take a while to get rid of...... so I then have a look at price movement usually on the higher timeframe (D1 chart) and identify where there is blue sky and no significant prior price obstacles. You will note on the D1 chart a high at around May which I want to be clear of for the H4 entry. That's why I set it there.......however remember this is not an exact science and there is no empirical support that prior price patterns such as prior support/resistance actually need to be observed in my approach. All you need to ensure is that the predictive entry is higher than the highest or lowest price of the current trend you are defining with the SDC and also higher/lower than the predictive entry of the lower timeframe for sequential trade orders. I actually think that it is best where possible to avoid your brain interpretating anything. As a result, so long as it is clear of the H1 predictive entry, that would in reality be good enough for me.

PS I re-read your query and note that you already using the back-test spreadsheet. Sorry about the confusion. Unfortunately given time constraints I haven't had the chance to go back to Jan 2013 and test out the EDTTI approach that uses multiple timeframes. It is a real pain in the neck and a slow process without an EA to do it for you. So we only have data for the multiple timeframe method since (late) Jan 2014 which I am forward testing after closing down the portolio early this year. To be honest with you, given my understanding of the technique now, I would bet my farm that the EDTTI will work on any instrument in trending periods and I know from the DTT testing that 2013 and 2014 on all instruments was a positive result.......however to be absolutely positive you would have to back-test yourself the EDTTI approach and see how it goes....if you want something to occupy yourself for the next eleventy million weekends. The alternative is to be patient and work in forward testing for the number of months necessary to convince you whether or not it works. :-)

The Recap continued.......

Greetings again all.....

Given that a number of participants are well ahead of the curve in their understanding this technique, rather than wait for next weekend to finish the recap, I thought I would give it a good nudge tonight to allow interested folk to fast-track in the learning curve.

Well since the last post of the recap we have had a few days to become familiar with:

1. the creation of a template for trading a single instrument using a clean chart;
2. applying the SDC with a standard setting of 1.0 to determine which timeframes are trending, their volatility and in what direction;
3. establishing predictive entry lines in the primary trend direction; and
4. a bit of a discussion as well from queries.

What you may note regarding the positioning of predictive entry lines in the direction of the primary trend is that you have effectively established an impermeable barrier on all timeframes that ensures that if price breaks out on the upside, that you will always catch the movement. This is unlike other traditional trend following models that are more selective in entry technique based on a particular trending pattern.

Of course this means that on average you will face more losing trades if price movement is of a random walk nature as opposed to a biased determinate pattern. For this reason, with this approach you must be aggressively risk averse...but with it comes the rewards if price indeed makes a protracted run in a single direction. In trending periods, this approach will literally clean up delivering exceptional returns for a trend following system, but in non-trending periods, you can be hampered by excessive drawdowns if you are not miserly with your risk management approach and if you do not set fairly wide predictive entries to allow a trend to form and discourage trade entries within potential price congestion ranges. Do not be concerned that you have left the run too late for trend entry with the placement of predictive entries that appear to be too far away, as all you are doing is adding substance to the possibility that the trend is a real trend by doing so.

Anyway, back to the charts.

Now the fun begins as we are now ready to trade......So lets step forward in time with our USDJPY template and see when our first order is triggered. But just note that this post is probably going to be the toughest to get a handle on as it involves some maths...but bear with it because it gets simpler after here.

The Chart below highlights how price on the M30 timeframe is now approaching the beige predictive entry of that timeframe and the consistent directionality in price movement appears to be trending. Our trigger fingers are getting itchy.

Note how there appears to be a fairly well defined and orderly trend in directional price movement in the M30 SDC channel to this point, which is making us confident that the entire price series may just have that small bias, suggestive that this may not just represent random walk in action but possibly part of a longer term directional move associated with some significant economic factors driving the MN primary trend.

Of course we never really know whether the trend is simply a random walk movement or random price movement with a small bias, but we like to think our chances are better in riding a portion of a larger secular trend of the market with this unfolding price action.

At this point in time when a potential trade is imminent we are checking on the nature of volatility of price in the SDC to ensure that upwards or downward price movement in the SDC is not being significantly biased by a few bars in the overall channel.......now all seems good.....so we ensure we set an alert at the predictive entry line (if you are trading by charts) or if you are not, then get the order ticket up and waiting.

What we are waiting for is the bar immediately prior to entry where we define position sizes and our initial stop for the trade and initiate a pending order or prepare for a manual trade entry. Now be aware that you can anticipate that the next bar is going to breach the predictive, but you can never be sure, so the following process may need to done a few times (adjusted) before an entry is confirmed.

So lets move forward again in time on the charts and see what the market brings.



Ok, we have a breach of the beige predictive entry line (M30) which is actually visually a bit late for the purposes of this point of discussion, but for demonstration purposes at least this breach has confirmed that you have entered. So lets imagine a point in time immediately prior to the bar before the breach of the predictive entry line.

Take the following steps to prepare the setup.

First you drill down into the relevant chart to clearly see whats going on.



Now where we are actually in time is on the bar immediately prior to the predictive entry breach but note the following process needs to be finalised on the *close* of this bar immediately prior to the bar of entry, so the final verdict of entry condition and position sizing is actually made when the new bar of entry commences forming.

Why?.....because if you extend the SDC to the current bar it continues to redraw every time price moves on that bar (on the current close) and we do not want an SDC that repaints.

So what info do I need to set the pending order entry? You need the price of entry (which is the value of the predictive entry line) and we want to ensure that if price movement goes against us that the maximum we lose is a defined trade risk % or less. Hence we want to define a stop that is consistent with this technique and not an arbitrary one and clearly defines when the approach tells us when its time to get out. Given that this approach measures a trend from where price action moves within the SDC channel, a suitable stop is therefore where price touches the outer SDC channel to ensure that the SDC itself is breached. Given the required location of the stop, we then want to ensure that our trade risk exposure is defined.

For the purposes of EDTT1 we use the following trade risk% guidelines:
M30 = 0.25%
H1 = 0.25%
H4 = 0.25%
D1 = 0.50%
W1 = 0.50%
MN = 0.50%

What does this mean as far as risk goes?......well if you have a $200K trading account, you are therefore only risking $500 for an M30, H1 or H4 trade or $1,000 for a D1, W1 or MN trade. If you dont compound profits, which I dont, your trade capital (provided you are not in a capital deficit....which has occurred once or twice with these various approaches...and in those instances I just top up the account using past profits)....will stay at a constant balance of $200K leading to a consistent $ risk on your trades.

If you compound profits (put them back into trade capital) then simply ensure these trade risk %s are observed and calculate the relevant $ risk each trade.

Why the different level of risk between M30, H1, H4 and D1, W1 and MN?.......This is because the back-test is concluding that trends on the longer timeframe have a greater directional extent. I have a hunch why that might be, but cannot with any certainty say exactly why. The hunch relates to the notion that sequential real economic underpinning factors are accentuating trend bias (or levy flight) on the longer timeframe. This is also supported by the research paper a few posts back examining trends over history and their nature. There has been an observed reduction in the number of trends of shorter duration over history but no significant change in frequency of long term trends has been identified.

But arent I exposed if I have trades on all timeframes running at once?.....No you arent, because given the staggered entry into trades, to have 6 trades on a single instrument running at once, you have built up a considerable on-screen profit on some of the timeframes which therefore more than compensates for any risk deficit on a single timeframe. What this effectively ensures is that we are risk management Nazis with this approach and find that we are never exposed for more than say 1.00% at any one moment in time. We also use a trailing stop once the trade has been entered (which well get to shortly) that ensures that our risk is actually even lower than these minimal levels.

But thats just crazy talk. You cant effectively trade with such low trade risk?.......thats where I beg to differ and is a singularly powerful myth that is perpetuated in trading circles and has been disproven by empirical testing with these approaches. I am not sure where the old 2% rule emanated from.....but that is far too generous with this technique if you want sustainable returns.

Ok now back to the last chart. We know we are applying a trade risk% of 0.25% for the M30 timeframe and we have an entry price of 102.45 determined by the predictive entry line, so what position size is required for the trade and where do we place our initial stop (which I refer to as the initial hard stop)? .......Well for this, we need to define the placement of the stop and the distance in points or pips between the stop and entry referred to as R.

Refer to the chart and look at the placement of the initial stop at the most distal end of the SDC. This is where you place the initial hard stop for if price reaches this point immediately after entry, this is where your SDC will be breached. The value of the stop on the chart above is therefore 101.822. Ok, now to calculate R which is the difference in pips or points between entry and stop = (102.45 101.822) = 0.628 = 628 points or 62.8 pips. The reason that we need to know R is that later in this technique, when the trade is 2.5R in profit (or 157 pips in profit) you will then need to alter the SDC width from 1.0 to 0.5 to reduce the width of the channel and restrict the distance to the revised outer SDC channel and ensure not too much profit is sacrificed when the trend bends.

So far so good. We have a graph entry price (E), A graph stop price (S) and now R. Now comes the harder part, we want to convert R into a Capital risk measure that incorporates an estimate of all costs of the entire transaction but still lie within the required 0.25% trade capital risk.

What do we mean by the entire costs of the trade transaction? This means that we need to incorporate into this risk measure:

1. the estimated spread of the entry transaction;
2. an estimate of any anticipated holding costs; and
3. a contingency estimate for slippage.

Now it is important to note the following feature when trading by charts. Typically you will be used to a buy/sell spread for a transaction, When trading by charts the midpoint between the selected buy/sell spread is used so we need to break the spread into two separate components which is approximately of the transactional spread costs.

We break this into two components being the spread on entry and the spread on close and remember that the spread throughout the transaction changes. Now in live trading you dont want to be stuffing around calculating the actual spread so I will show you how I estimate a number which I use as a convenient way to quickly ensure I have conservatively estimated total transaction costs.

To estimate an adjusted spread on entry and an adjusted spread on close that includes an estimate of holding cost and slippage contingency, you need to simply adopt a very conservative estimate that more than compensates for these necessary costs incurred in trading. I take the average spread and halve it and then adjust this entry and exit adjusted spread calc x 1.2. In my instance using IG markets, the average spread plus these estimates create an adjusted entry and exit spread of approx 30 points for entry for forex and spot commodities and 10 points for indices. I am conservative by nature and these estimates I find more than compensate the trade transactional costs, but you will need to estimate these yourselves dependent on your broker.

Once this is considered we are then in a position to confidently and conservatively estimate position size.

So we need a formula here. I actually build this formula into my database where I record my trades so it is automatically calculated when the inputs are provided.

Here t'is

Rec # Lots: -Round(([Trade Risk %]*[Opening Balance])/([Risk Pts]*[$ Per pt per 1Lot]*[Lot Size]),0)

Lets see what we can define here:

1. Round = Simply rounds the result (number of Lots);
2. Trade Risk % = In this instance = 0.025%
3. OpeningBalance = $200,000
4. RiskPts = We need to define this separately (refer to next formula).
5. $ Per pt per 1Lot = $1.00 for all instruments
6. Lot Size= 0.01 for microcontracts, 0.1 for mini contracts or 1 for full contracts

So now to define RiskPts

Risk Pts: IIf([Long_Yes]=-1,Round((([Entry Stop]-[Entry Price])*[Spread Pts Conversion]),0),Round(([Entry Price]-[Entry Stop])*[Spread Pts Conversion],0))

Dont be scared of a conditional if statement. All this does is create a formula that relates to whether the trade is a long trade or a short trade.

1. We also round this entire result to 0 decimal places.
2. Entry Stop = Graph Stop Adjusted Entry Spread = 101.822 0.03 = 101.792
3. Entry Price = Graph Entry Price plus adjusted spread = 102.45 + 0.03 = 102.48
4. Spread Pts Conversion is a conversion ratio that is applied to the instrument to convert say 30 points to 0.03 in decimal notation.

Below is a Spread Pts Conversion Table for the instruments I trade.

.....anyway enough with the formulas.

The end result of all these calculations is a conservative recommended position sizing of 73 micro-lots. If the trade moves a distance R, then that brings the total loss to <= 0.25% of trade capital or <=$500.

..well that was a mouthful so I will show the results I enter into my database and give you reasons for the cost estimates I include in my risk assessment.



I use a tailored Access database (dbase) to record trade details and you are also going to need something similar such as a database or spreadsheet as quick calculation and recording of trades is essential.

In my trade display I have:
1) An opening balance (OBal) reflective of my trade capital account balance (eg. $200K);
2) Details of the instrument (Pair) I am trading....in this case USDJPY;
3) The trade risk % = 0.25%;
4) If the trade is long or short (L=1), where long is a tick and a short trade is blank;
5) Disregard the countertrend column for the EDTT1 as we only trade in the primary trend direction with this technique;
6) The Time frame of the trade = M30;
7) The graph entry price = the price of the predictive entry line = 102.45;
8) The entry date= 30/7/2014;
9) The entry time = 13:30 which is GMT+10 in my zone;
10) The buy/sell adjusted spread for entry that I estimate at 30 points for forex as detailed previously;
11) The graph hard stop = 101.822
12) And then the dbase calculates by formula an estimate of position size = 73 in microlots;
13) The actual position size taken for the order = 73 microlots which is the same as the recommended column; and
14) The dbase calculates a result of 2.5R (2.5 x R) which provides information later to us that we use when deep in profit to reduce the SDC to 0.5 width (more on that later).

Ok....now note that we cannot complete trade close details including the adjusted close spread as that lies in the future of this trade and note also that we have no defined profit target as we let the profits run.

So we have all the information we need to make the trade......I therefore buy 73 micros or 7.3 minis when the predictive line is breached and know that I am risking worst case $500 bucks.

Now remember as discussed before that the placement of the initial hard stop is at a specific location where according to this technique, thats the point you must get out of the trade if the trade immediately reverses in the same bar to touch this hard stop.......but on the following bar and thereafter the initial hard stop is replaced by a trailing stop that sits on the SDC outer channel and progressively follows price as the SDC breathes to the tune of the market (more on this later).

......so what we therefore need to do to create the trailing stop is superimpose a hashed line over the stop line of the SDC so that we can set an alert or pending stop on that line to automate exit so we dont have to worry further about it. Refer to the chart below to see this hashed superimposed line on the outer SDC channel.



Just a note about the initial hard stop. While we initially set a hard stop and then replace this with a trailing stop, I still retain the hard stop as a contingency buffer (or secondary stop) just in case we have a fast price movement or extraordinary event on our hands where the rapidity of movement just flies through our trailing stop. This is just the a secondary precaution measure and you might want to do the same thing.

Now the last thing we do before moving on following an entry is to change the nature of the predictive entry line to a new format as shown in the prior chart above to visually reflect that a trade is now active.

At last, we have completed our trade management chores and now it is up to the SDC to take trade management over...... So we have a trade in play.....and we move on in time which will be the subject of the next post.

C

The Recap continued.......

Ok we have come a long way so far and not much further to go.

So now we have now stepped forward in time as detailed in the chart below.



The trade on the M30 timeframe is progressing well but you can see that price on the H1 timeframe is approaching the predictive entry line. Once again the trigger finger is getting itchy and we follow exactly the same process outlined in the prior post.

But before we get there, we need to discuss what we do to the SDC on the M30 timeframe when it reaches a new high for a long or new low for a short. We need to redraw the SDC each time it does this. This ensures that the additional price data is included in the SDC channel which allows it to breath like a Bollinger Band.

Given that we redraw the SDC, we therefore also need to redraw the hashed trailing stop line over the new SDC outer channel and re-set our alerts and pending stop orders.



Now the process undertaken of redrawing an SDC is a regular one when new highs or new lows are made and I always keep the latest SDC channel on the chart but delete prior SDC channels EXCEPT the original SDC channel that was established in entry. The reason I keep the original SDC is simply as a reference tool as it contains important info such as the initial hard stop position and the trend slope and width immediately prior to entry.

Now don’t get too anal about the redraw every single time a new high or low is reached, but it is good practice to keep the SDC current. Note that in the chart above, the redrawn channel commences at the same location as the original SDC but now extend to the new high. Note how there has been an increase in width and slope based on the additional price information that is now contained in the channel. The width setting is still 1.0 but increased volatility has naturally widened the channel. Furthermore the trend channel becomes steeper given the price movement in your favour which means that the trailing stop accelerates towards a breakeven position.....oh the beauty of the SDC in naturally managing trades J

...but what’s that?...is there a breach on the H1 timeframe? Have we overshot the mark and shouldn’t we have already taken a new trade?.....Great observation. You are quite correct so let’s back track just a bit.

Now we have a predictive entry touch on H1 but given the rules of the prior post we were ready for this and immediately on the bar of entry we drilled down to the H1 chart and followed the same process as before.



The result of the trade entry on H1 are detailed on the chart above and the records of the trade taken are reflected in the following database snapshot.



Ok so now we move on again in time.



Damn...it looks like the trades on the M30 and H1 timeframes have been stopped out!!!!!.

In my next post I will describe what we do in exiting positions.

C

The Recap continued.......

Almost there guys.

This post will focus on what we do when we exit a position.

So refer back to the current template.



When the alerts sound or pending stops trigger we drill down into the relevant chart to see what’s going on.



Now note that we have not been able to further redraw the SDC from the high on the 30 Jul as no higher high from that point on occurred. That therefore effectively locked the SDC in place until price ultimately decided to touch the hashed trailing stop line. On the touch we are out of the trade which is preferably done with an automated order (which is made available with ProRealTime’s trade by charts functionality).

Given that our exit was below our entry we are going to cop a loss but this will be far less that the original 0.25% of trade capital as over time, the trailing stop has faithfully mirrored the extension of the outer channel of the SDC which defines the trend boundary.

For example, the trade result is detailed below as a loss of ($71.54) or -0.14R which is far lower than the initial hard stop of 1R or a loss of $500. Refer to the closed trade details captured by the dbase for M30 below.



.....but a loss is a loss and expect many of them...Just suck it in and remain confident in the approach.

Once a trade has exited it is time to clean up the chart by removing unnecessary information associated with the prior closed trade and by resetting the predictive entry as detailed below.



Following the cleanup, then we can return to the template.

But wait....we have also been stopped out on the H1 timeframe as well as the alerts are sounding.





Ouch that one hurt a bit more.......once again just suck it in and reset the chart.



.......Ok you now should start getting the picture so what I will speed things up in the next post and just roll through time to show the charts and the steps taken.

C