Diversified Trend Trading Approach

Improving the BB Mean Reversion System

Below is the result of a slight variant of the BB mean reversion tests described in post 719 where I have simply adjusted the SMA trend filter from 200 to 100 to increase trade frequency yet ensure that the BB system is only turned on when markets are in trending phase.

The mean reversion system is proving to be a very nice addition to the portfolio system suite due to it's lower volatility of returns. It might not have the punching power of the Donchian breakout systems, but it is proving to be quite reliable across my investment universe.

B3C Unscaled Gross Result



B3C Scaled Net Result with Profits Reinvested



DAll MT/ FUSEX/ B3C Blend
The best of the blends so far. The Annualised Sharpe ratio is continuing to improve as is CAGR and CAGR/Drawdown.



Regards

C

Food for Thought in Relation to Trend Following Versus Mean Reverting Strategies

A very good article from Laurent Bernut that is well worthwhile spending some time digesting.

Regards

C

Music to back-test to...... :-)

Dual Moving Average Crossover Systems

I have been turning my attention to Wisdom's Dual Moving Average Crossover Systems and seeing how they fly over my portfolio universe. As Bassramy (Bass) has pointed out, the profit contributions from the DMA systems for the Wisdom Portfolio have been significant. For example, here are January 2016 results for the portfolio.

You can see from the profit attribution chart below, for the last 12 months, the DMA-L system has been the 2nd greatest profit contributor to the system suite. The DMA-M and the DMA-S systems however were negative profit contributors for the period.



....anyway, I thought I would see how my portfolio stacked up using these systems for the period 1 Jan 2000 to 31 Jan 2016.

The 3 DMA systems are configured as follows:

 

1. DMA-S - A short term trend following system - Uses a short term SMA of 10 periods and a longer term SMA of 20 periods. There are no stop losses engaged by this system. You enter when short term SMA crosses long term SMA (long or short) and exit when short term SMA crosses long term SMA. You are therefore in the market all the time. ATR is used for position sizing only. Settings for the ATR for DMA-S are 39 (days)/2 x multiplier).
2. DMA-M - A medium term trend following system - Uses a short term SMA of 20 periods and a longer term SMA of 50 periods. There are no stop losses engaged by this system. You enter when short term SMA crosses long term SMA (long or short) and exit when short term SMA crosses long term SMA. You are therefore in the market all the time. ATR is used for position sizing only. Settings for the ATR for DMA-S are 99 (days)/3 x multiplier).
3. DMA-L - A long term trend following system - Uses a short term SMA of 50 periods and a longer term SMA of 200 periods. There are no stop losses engaged by this system. You enter when short term SMA crosses long term SMA (long or short) and exit when short term SMA crosses long term SMA. You are therefore in the market all the time. ATR is used for position sizing only. Settings for the ATR for DMA-S are 399 (days)/5 x multiplier).

Well....unfortunately the results were shockers for the my portfolio. Below are results of the un-scaled performance results before any fees.

DMA-S

Notes: Less that half the markets in the portfolio were profitable. Soft commodities and metals performed well but the forex and indices were eaten alive. Best results of course were seen with the recognised trending instruments over the period like EURUSD, BRENT OIL, SUGAR#11, COPPER & SILVER. What nailed this strategy was the whipsaws and frequent sideways congesting markets over the period using the associated system settings. While it may have been a darling for the Wisdom universe, it would not be a strategy I would be considering for inclusion in my portfolio due to it's extreme variability of returns which can be simply eye-balled on the equity curve. No need to progress further with this one. If there was any hope for this system, you would need to consider including a filter to prevent trading through noise and push trade entries into clear blue sky at the distal extremes of average volatility. In a nutshell there is no clear edge in this system applied to my universe, but there are periods of powerful returns which might lead the untrained eye to conclude otherwise.


DMA-M


Notes: Eeeeeekkkk. Not much to really say at all apart from ....keep away!!!! We clearly need to move to higher timeframes to get clear of the noise. I certainly aren't getting similar results to Wisdom, and this is either contributed to the difference in portfolio universe or else I have missed some secret sauce in Widsom's system definition. Things started going really pear shaped following 2008.


DMA-L


Notes: Slightly better but no cigar. As I suspected as we stretch out into longer term horizons the influence of noise is reduced and some of the more robust trends of significance are captured. 24 instruments were successful while 17 were not. We are entering a sweet spot but the volatility of returns still does not give much room to move if you want to scale it up or include it in the portfolio suite we have racked up so far.

General Notes: I have long been suspect of systems that are in the market at all times. While a successful strategy up to 2008, during periods of extensive 'non-trending', these strategies get eaten alive. It is necessary to make adjustments to this style of strategy to include filters that avoid this noise, but in doing so, you fundamentally alter the robust nature of the strategy. In essence, the markets have not been favourably responding to this style of system so rather than tweak the system, it is necessary to look to new pastures and identify those systems that are fundamentally correlated with the market conditions. What is clear so far is that momentum based breakout systems such as the Donchian breakout suite of system are better suited to my investment universe than traditional trend following systems.


PS Bass I will move onto the TMA systems of Wisdom and see if they do any better, but am already suspect that the Wisdom systems are more applicable to the futures space than my investment universe. I will keep you posted.

Regards

C

State of the Trend - Discretionary Systems - DTT, EDTT and EDTT1

I thought I might go back a bit and post the state of the trend for the discretionary systems described in this thread for those of you who are still interested in these systems.

Rather than describe the charts, I will let you guys make your own interpretations. In general however not much happening on any timeframe for the universe these systems are trading apart from say GBPSGD, GBPUSD, USDCAD and Gold.

Enjoy with a bit of the mandatory music.



SPI200 (ASX200)


AUDCAD

Attached Image (click to enlarge)

AUDJPY

Attached Image (click to enlarge)

AUDUSD

Attached Image (click to enlarge)

CADJPY

Attached Image (click to enlarge)

DAX30

Attached Image (click to enlarge)

DJI30


EURAUD


EURCAD


more to come.

C

State of the Trend - Discretionary Systems - DTT, EDTT and EDTT1 continued.....

EURJPY

Attached Image (click to enlarge)

FTSE100

Attached Image (click to enlarge)

GBPSGD

Attached Image (click to enlarge)

GBPUSD

Attached Image (click to enlarge)

NIKKEI225


NZDCAD

Attached Image (click to enlarge)

NZDJPY

Attached Image (click to enlarge)

NZDUSD

Attached Image (click to enlarge)

More to come

C

State of the Trend - Discretionary Systems - DTT, EDTT and EDTT1 continued.....

USDCAD


USDCHF

Attached Image (click to enlarge)

USDJPY

Attached Image (click to enlarge)

USDSGD

Attached Image (click to enlarge)

Silver

Attached Image (click to enlarge)

Gold

Attached Image (click to enlarge)

Status of the Portfolio - Unrealised profit as at 27 Feb 2016

Attached Image (click to enlarge)

That's it

Have a great weekend all. :-)

C

Very nice results C. Well done

Cheers NT. Things are drying up a bit for the trends. Bring back early January I say . That's when things were cookin' with gas.

Thank you mate, not the results that we hoped for, still the TMA's to test.
Included some optimized parameters, portfolio Metrics with optimized MA Crossover results on a small portfolio of Futures with 10 years data testing period.

Mate. Those win rates for a MA cross system blow me away and great profit factors as well. Just a quick question? Were they in the market all the time, were stops or exits used and were any filters used?

Cheers Bass

Nice trading mate.. Hopefully it carries on for you ..


Its been an interesting year so far..

Thanks Ja

You have some nice instruments in your portfolio mate.

Feb has been a tough month so far mate and despite the unrealised sitting in the portfolio, it looks like it will be a marginally negative month for the discretionary portfolio dependent on what Monday brings.

Thanks for Bebe....awesome track...such a smoking voice. :-)

PS Your You Tube link has led me to a great place Ja.

Nice....You tube is great place isnt it...

It sure is, especially with a free mp3 converter

Groove Armada back in 2000....pretty impressive :-)

Lol.. I have one too.

YT - especially when doing Weekend spread sheets

GA is lined up next..

One of my favourite GA tunes..

Loving the music mate. It's hard to go past GA.......what consummate professionals.

Sounds like we all are deep in crunching mode this weekend. Music helps to pass the time..that's for sure.

You know, after all this back-testing I am starting to feel like a market historian. I can definately detect a shift in market conditions relating to trend following since 2000 and can now understand what some of the CTA's are talking about. Neils Karstrup Larsen from Top Traders Unplugged was talking about this with Bill Dreiss that there has been a notable change in the market since 2000 and in particular following 2010 and I am definately seeing this in my testing.

Traditional trend following technique clearly needs to shift out to longer timeframes to find alpha and the medium term/short term space is now a zone for momentum based breakout systems. It has been a notable feature of the equities market and the forex market and now appears to be infecting the futures space as well. While a generality, it is being detected across most instruments I am testing using either traditional trend following systems or momentum breakout based systems. I am assuming the major factors that might be influencing this are:
1) Increasing numbers of systematic trend followers in the market today that is diluting available alpha;
2) Increasing number of algorithmic traders and HFT which is fundamentally disturbing traditional market volatilitiy; and/or
3) Increasing central bank intervention which is preventing excessive volatility that is the bread and butter of trend following.


.....anyway, I will take off my historians hat and get back to testing. Should have the Triple Moving Average (TMA) tests out shortly. Not much joy on this front like the DMA tests.

C

I really notice it too.. If nothing else all this BT work gives great transparency into the Financial universe... It also reveals some things I really wasn't expecting..
its very interesting..

Neils Interviews are cool.. I have a constant flow of his podcasts when I'm driving..

Triple Moving Average Crossover Systems

This post focuses on a Triple Moving Average Crossover System that features in the Wisdom Trading arsenal and is described below.




Well....like the DMA tests of post 724, there is not much to see here. You have to step out to the longer term time-frames to start seeing these systems perform.

TMA-S



Notes: Like the DMA-S, less that half the markets in the portfolio were profitable. Once again, the short term settings were not sufficient to escape the noise of the short term and despite some great profit runs, the fruits of these profits were eaten alive by many consecutive losses.


TMA-M



Notes: A slightly improved result, but nothing to crow about.

TMA-L



Notes: A better result but certainly not a contender for addition to the current portfolio suite of systems. Simply too much volatility in the distribution of returns to have any appeal.

Well Bass, looks like I need to be investing a bit more time in momentum breakout systems and mean reversion systems. We have a couple of good ones up our sleeve already, so I will start extending my search for a few more good inclusions and maybe as a new project start trialing some of Gary Antonacci's systems in the forex, index and futures space.

Cheers guys...signing off for the weekend.


C