Diversified Trend Trading Approach

I hear you A....and deep down I know trends are just as valid as they ever were.....however it pays to be humble in this game. What I always snigger at on this forum is that we are arrogant creatures and are always thinking that we are reinventing the wheel....when in reality, what is valid today is just as valid as perhaps it ever was. The tools of the game have changed however but when brittle failure occurs in market structure, as it inevitably will....then trends will remain supreme.

Trend trading does require one to adopt a contrarian set of beliefs.

Anyone who attempts to explain why trends exist or persist will not arrive at a definite answer. To me, self-sustaining price movements are a natural consequence of speculative markets. The only way to ensure that trends will never again occur is to close down markets are ban people from trading.

I believe as long as the long term primary drivers for market movements are influenced by human intervention with global growth the machines will take a back seat.. but, never say never..

Interesting read C . . . thanks

You are probably right mate. I was talking to a drover the other day *literally* from Camooweal in North West Queensland who was lamenting about the days before long haul transport when they ruled supreme (and I thought I was in a time warp) so as you say, nothing is written in stone.

PS.....that Degiheugi was a great piece mate. Never heard of them before. I think I need to speak with Uncle Torrence and look into it further. :-)


EDIT........ how did they escape me? Luv it.






Cheers

C

The only enduring constant in this universe is Chuck Norris and we all know it. There will be the occasional cheap imitations like Van Damme, Stallone and Schwarzenegger and of course the cybernetic algorithmic versions, but Chuck is here to stay.

For all those who stray from the forum and actually have a life over the vacation period........have a great Xmas.

Guys....sorry about the deletion of the post. Unfortunately this method requires the release of specific EA's which is problematic due to permissions required and potential confidentiality concerns.

C

I have no life. I willing gave it all away. When you speak of diversified trends, how do you manage correlation?

Hi J

Trading multiple timeframes gives me some degree of reduced correlation, but there is no getting round the fact that medium to long term trend trading actually relies on correlation. For example most of the alpha achieved in trend trading occurs during market crisis periods such as 2008, where every instrument seems to be correlated. A tight trailing stop and no profit target is the best solution for this dilemma.

I personally think that we can sometimes be too concerned with correlation as a trend trader where we can miss some great trend trades, so I ring fence all timeframes and all instruments with pending orders outside current volatility so if a trend commences in earnest I am onto it without missing a single substantive move. However.....by trading multiple timeframes I am effectively scaling into a trend in a staged way and at each step of the way ensuring the prior trade taken on a trend has reached breakeven before I take a new trade on a different timeframe. This way if it goes pear shaped my portfolio is not overly exposed.

For example, a trend usually is first seen on M30 where I take the trade and wait until breakeven has been reached before I move on to the H1 timeframe to see if the trend is valid on that timeframe and so on and so forth up the timeframes. What this therefore achieves is that the shorter timeframes capture the swings of longer trends captured on the longer timeframes. In effect we are combining swing trading with trend trading that helps to a degree in flattenning the equity curve and assisting cashflow.

For example at the moment my active trades are dominated by CAD pairs which are all highly correlated but deep in profit with a tight trailing stop on each timeframe...so the portfolio is not exposed and I have not taken a selective stance in which CAD pair I trade. Some of the CAD pairs are trading on 6 timeframes the majority of which are deep in profit. An example is provided below of NZDCAD where M30, H1, H4 are deep in profit (no portfolio exposure as each trade's worst case is well above BE) and have recently entered D1 with W1 and MN pending.



Trade risk is very low per trade (approx 0.25% to a max of 0.5% of equity) to allow for protracted non-trending periods without getting into deep drawdown territory. Each loss is a gnat bite and inconsequential to me allowing me to focus on the next thousand trades which is necessary if you pursue this approach. It is not a cashflow solution but rather a long term wealth creation solution which is not everybody's cup of tea.

In essence, when the trends occur you are into them, but to survive the protracted periods where trends are not prevalent across the portfolio, keep trade risk % very low and do not over-trade at the first sign of a trend.

Of course the equity curve for the portfolio has it's volatile profile which is typical for a trend trader. If you stick simply to trends, then there is no way around this feature. That's why I was interested in your technique (trend trading with Bollinger reversion using ADX filter) and also having a look at other ways to achieve system diversification to flatten the equity curve and make more productive use of trade capital using correlation as a guiding measure.

I hope this makes sense J. :-)

Lol... Chuck is the man..

Merry Xmas C..

Monthly Performance Metrics of the Top Established Trend Following Funds

Greetings,

I hope you all have had a great Christmas and are recharging after a challenging year in trading. I thought as a regular monthly inclusion to this thread I would monitor the progress of the 'best of the best' in the fund management space that choose trend following as their dominant modus operandi.

The following performance metrics (shown in the next few posts) reflect the performance of 47 large fund managers who have been operating at least since 1 Jan 2000. Many of these funds have been in operation since the early 1980's but to be able to monitor comparative performance between these funds, I have chosen the start month for the performance monitoring as 1 Jan 2000.

For the purposes of the monthly update I will continue to use these 47 FM's and see how survivorship bias affects the results. In the future I may augment the results to include additional FM's that also have an extensive history of performance, but at least for the moment the base month for comparison will be set at 1 Jan 2000 to make things simpler in the analysis.

There is a deliberate reason for my choice of start date. As many of you are aware, the performance results of trend following often tend to be 'gloriously misinterpreted' and can be misleading to many that are testing their own trend following systems and wonder why their results are so dismal when compared to the big guys. I engage in extensive back-testing of trend following strategies as you may, or may not be aware, and due to the limited data available to me across a wide investment universe, the best I can achieve is a data set that extends back to 1 Jan 2000. Unfortunately however, trend following has been a tough game since 1 Jan 2000 and we have not enjoyed the heady returns associated with the 2 decades prior to this date. This is not to say that those level of returns won't again materialise at some time in the future, but by and large, since 1 Jan 2000, performance metrics of trend following as a class has been subdued. There have been some windfall years during this period but recent action at least since 2009 has put pressure on this trading class.

I have created a hypothetical Trend Following index (TF Index) that consolidates these results and is used as a basis to determine the general health of trend following as a composite. This will be our general guide for the performance of trend following as a distinct trading methodology.

Further down the track in the next few months or so I will be using this data as a basis to demonstrate to you all how you can construct a powerful and robust portfolio using a range of effective selection measures that flatten the equity curve and allow you to outperform the segment as a 'Fund of Funds' solution in the trend following space. If you are really after wealth creation, doesn't it make sense to use the best of the best and stand on their shoulders. Obviously the minimum account sizes and investor type restrictions required to invest in these big boys is an impossible obstacle to hurdle for the meek retail investor, however the principle can be applied to any fund of fund solution, so why not in principle demonstrate the theory with these big boys anyway. The same principle can be applied to the best retail funds out their.

So since this is the first post in this series, I will include some explanatory commentary for the attachments (and inserts) so we all can be familiar with the analysis, but in subsequent monthly updates, I will simply include the attachments with little commentary to save myself time.

Here we go..................but.........but..............you know I have to include some music as well. Next post we will start. :-)

Regards

C

November 2015 Performance Metrics of Established FM's in the Trend Following Space



Explanatory Items/ Definitions:

Strategy = A summary descriptor of the trading strategies employed by the fund.

1. Div - Diversified
2. Syst/Disc = Systematic or Discretionary (and percentage allocation if relevant)
3. TF = Trend following
4. CF = Counter-trend
5. Mom = Momentum
6. Tech = Technical
7. Patt = Pattern recognition
8. Cyclic = Cyclic and/or seasonal
9. Spread/Hedge = Spread or hedge techniques

CAGR = Compound Annual Growth Rate (or Return on investment since 1 Jan 2000)
Rank = Performance ranking of fund in association with CAGR. Yellow represents top 10 performers by ranking. Red represents bottom ten performers by ranking.

Max Draw% = Maximum Drawdown in realised equity since 1 Jan 2000.
Rank = Performance ranking of fund in association with drawdown%. Yellow represents top 10 performers by ranking with lowest drawdowns. Red represents bottom ten performers by ranking with highest drawdowns.

Adjusted CAGR (Scaled CAGR) = Compound Annual Growth Rate (or Return on investment since 1 Jan 2000) assuming performance scaled to a maximum drawdown tolerance of 20%. Adjusted CAGR represents an ideal performance measure as it rates performance both in terms of total return and drawdown exposure.
Rank = Performance ranking of fund in association with Adjusted CAGR. Yellow represents top 10 performers by ranking. Red represents bottom ten performers by ranking.

S&P Correl = Correlation of fund return performance against the S&P 500 Total Return index (adjusted to include dividend and splits). A correlation of 1 represents perfect positive correlation. A correlation of -1 represents perfect inverse correlation. This performance metric is included as it demonstrates how trend following as a class has very low correlation against the S&P 500 Index and is an ideal class of systems to include in any diversified portfolio.

Skew = Measures the asymmetry of the return distribution. Trend following as a class has a unique skew profile as the majority of trading systems have negative skew meaning that performance is characterized by many small wins with the occasional large loss. Trend following on the other hand exhibits positive skewness meaning many small losses with the occasional big win.

Kurtosis = Measures the tailedness of a probability distribution and is a descriptor of the shape of the probability distribution of returns. Higher kurtosis means more of the variance is the result of infrequent extreme deviations as opposed to frequent modestly sized deviations. Trend following as a whole performs best during crisis events where wins are extreme and as a result typically reflects higher positive kurtosis.

Sharpe = Annualised Sharpe Ratio. Sharpe ratios reflect the variability of excess returns above/ below the risk free rate. Note that the relatively low sharpe ratio's demonstrate that there is not much alpha in the market since 1 Jan 2000 for the trend trader.
Rank = Performance ranking of fund in association with the Sharpe ratio. Yellow represents top 10 performers by ranking. Red represents bottom ten performers by ranking.

Nov/2015 return% = The monthly performance return for the fund. Obviously as a trend follower you are less concerned with these monthly results and more concerned with the long term result, however they are useful in demonstrating whether the particular month was trending or not. Note however that monthly performance reflects closed trades hence large results in a particular month may mean that the trends are ending as positions are being exited and profits realised.
Rank = Performance ranking of fund in association with the monthly performance. Yellow represents top 10 performers by ranking. Red represents bottom ten performers by ranking.

Last 12 Months return% = The rolling 12 month performance of the Fund to the current month. A good guide that demonstrates recent performance.
Rank = Performance ranking of fund in association with the 12 month rolling performance. Yellow represents top 10 performers by ranking. Red represents bottom ten performers by ranking.

TF Index = A simulated theoretical index that averages performance of the entire suite of FM's into a consolidated result.

S&P 500 TR = The S&P 500 Total return index

November 2015 Performance Metrics of Established FM's in the Trend Following Space

TF Index Benchmark Performance
I want to closely examine the performance of the TF Index as it demonstrates the power of additional diversification in enhancing performance metrics. While it may never be the best performer of the entire fund listing, principles of diversification through consolidation allow it to post some serious results that in theory should keep it in the strong performance category. Remember with 47 funds to choose from, how are you sure you are going to select the winner. A Fund of Funds solution is therefore designed to achieve the best that can be delivered by all funds. As a Fund of Funds, you can get these benefits but is essential that your index must be equally weighted or better still volatility/drawdown weighted and not hamstrung by weighting by Assets Under Management (AUM) like typical indices.

As suggested, the TF Index is a strong performer of the group being the 7th highest ranked performer against adjusted CAGR (my preferred performance metric). It also ranked as one of the lowest drawdown options for the period and had the 10th highest annualised Sharpe ratio. So far so good for the Index.

Later on in this thread we will visit some options for consolidating portfolios into various Fund of Funds and we will explore principles of periodic re balancing and various techniques and the effect it has on accelerating performance returns towards the sky. So stay tuned to this space.



The table of annual performance and equity curve of the TF Index demonstrates what a tough game the trend followers have had in recent years. Theories abound for this subdued performance subsequent to the GFC (such as central bank intervention etc.), however whatever reason that may be inferred, clearly 2011 to 2013 were tough years for the trend traders. Most of the Funds experienced their worst drawdowns during this period but fortunately 2014 managed to pull the trend traders out of the slump.

Despite the relative lackluster performance since 2000, trend following has clearly outclassed the returns of the S&P 500. Also the maximum drawdown of the TF Index since 2000 of 18.11% is far superior to the max drawdown of the S&P 500 TR Index over the same period of a nail-biting 51% (refer to prior performance metric table).

The Investment Context - S&P 500 TR Performance against Risk Free Return



The chart above compares the performance of the S&P 500 TR Index against the risk free return, which for the purposes of this assessment and for Sharpe ratio calculations is the 3 month Treasury Bill.

Equities in general have had a horrid performance over the period from 1 Jan 2000 where they under-performed the risk free rate and only exceeded the risk free threshold in 2013. Investing in the S&P 500 would have experienced torrid drawdowns in both 2002 and 2009 reaching levels in excess of 50%. Who said 'buy and hold' and passive management is a prudent strategy? If there is any reason to seriously consider investing in the CTA space, then this is it.

November 2015 Performance Metrics of Established FM's in the Trend Following Space

Best/ Worst Performers for the Monitoring Period

















That's it for the month guys.

Enjoy.

C

Fund 2: 74.76% Drawdown
Fund 3: 91.46% Drawdown

D:

I wonder if being discretionary had anything to do with this.

LOL. Both are 100% discretionary and I get your drift. As I was pulling the data together and going through the funds (about 600 of them available on IASG across all style of Funds) after a while I just needed to look at their trading style and it's strategic composition to get an idea how they would perform in the long term.

There is no question that long term systematic trend following techniques make up the lions share of alpha in fund performance at least over the last 20 or so years.

Imagine having a drawdown of >90%!!!! When is enough, enough for gods sake. I know Bill Dunn has balls of steel with 50% drawdowns as his tolerance range, but surely after that you must start to get an idea that your systems employed are simply not sustainable.

Both those Funds had wicked accelerating performance up to 2008 and would have been the darlings of the market. As with nearly every example I come across, those that have a rampaging equity curve ultimately come a cropper. It all ended in tears after 2008 and they were never able to recover. Despite their rhetoric on prudent risk management.....not prudent enough so it seems.





Having a read of their investment strategy is like navigating through a tomb of economic white papers. They are adamant FA analysts. Perhaps the KISS principle might have worked better where they simply hopped on board the trends.

I also notice that they are the only discretionary funds in the list. lol.

With something as simple as systematic trend following im actually puzzled as to why there are so few prominent players.

To be fair, there are many more systematic trend following fund managers but I was only after the ones that existed prior to 1 Jan 2000 as I want long term historical performance data for what is coming up. Interestingly from 2000 onwards there were a large number of new entrants into that space.....probably it finally sank in. I never said Fund managers were smart and quick to adapt. There are a billion 'crusty and old' buy and hold FM's still out there for example.

Any idea how many new entrants are left? I have a feeling that new players won't stick around for long.

Roughly from the list I was looking at probably around 50 or so new entrants specifically in the systematic diversified trend following space that are still around today, but not sure if any more new entries have simply vanished over this intervening period.....and the list is simply those that are tracked by IASG. Also some are simply new programs from existing established managers. Many are starting to get their teeth into pattern recognition in conjunction with TF and this seems to be the new kid on the block that is getting attention lately.

I tend to agree with you that most will not have the patience that trend following requires and therefore leave the space right at the time trend following takes off again.....working on Murphy's Law. :-)

Change to Monthly Performance Monitoring of DTT, EDTT and EDTT1

Guys,

To make things easier for me, I am going to recast monthly performance monitoring metrics of my discretionary systems described in this thread (DTT, EDTT and EDTT1) in a similar format to that I am now taking for the proposed monthly fund management reporting detailed in the most recent posts above.

It saves a lot of time for me in the crunching but there are some things to be aware of.

1. I will only be reflecting performance results since 1 Jan 2015 for performance reporting of the DTT, EDTT and EDTT1 as prior to this date, I was only testing the DTT and to keep apples as apples and to more accurately describe comparative performance for those interested in these techniques, then I have excluded the historical results of the DTT prior to this date;
2. Unfortunately we cannot directly compare the results of my discretionary approaches against the FM's as they are materially different being Gross performance results as opposed to Net performance after deducting management fees from the result. Despite this issue, I have included the Index results in the equity curves so we can see how we are tracking against the overall TF Index and the S&P 500 TR Index; and
3. The analysis will therefore only compare relative performance between these approaches.

Below is the new proposed format for which the Monthly analysis of the DTT, EDTT and EDTT1 will take. I apologise if this inconveniences any forumites who are tracking these approaches, but the time saving to me in adopting this new method is significant. Slight differences will be observed in drawdown method as the approach that I now adopt is consistent with the methods reported for the FM's.

Regards

C