Diversified Trend Trading Approach | Page 82 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is Grid Trading (combined trend following and counter trend) Profitable? 3 replies

Making a good diversified portfolio 10 replies

Diversified EA Money Management Code 10 replies

HAMA PAD - A Simple Trading Approach 7 replies

Trading a (semi) automatic approach...but want more 11 replies

Attachments: Diversified Trend Trading Approach
Exit Attachments

Diversified Trend Trading Approach

  • #1,621
  • Jul 11, 2017 8:08am Jul 11, 2017 8:08am
Quoting Copernicus
{quote} You have a point R. Sun yourself on the beach and periodically rotate your portfolio as opposed to sweating at your trading desk developing things from scratch.....mmmmmm.....let me think about that for the moment.......I think you win Here is a tongue and cheek post I did that showcases how fund rotation using risk weighted performance measures for filtering options can create impressive performance metrics. Just note that the minimum investments required...
Please note I wrote about ETF/mutual funds where minimum investment is 30 USD. (At least in Poland)....

and if you would chose correct ETFs then such are the 1 year results:

Attached Image (click to enlarge)
Click to Enlarge Name: 5psZ5AX.png Size: 207 KB
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,622
  • Jul 12, 2017 4:06am Jul 12, 2017 4:06am
Quoting Copernicus
{quote} The art lies in delivering sustainable risk-weighted returns. This falls on deaf ears in this forum.
This forum is made to attract people to talk about fx/strats/everything&nothing in order to hike up on alexa website rank for public exposure, to be able to sell publicity space/banners to brokers.

Some traders here are only for fun, some are here to attact at their end newbies into a dark managed account or signal services, and some to join at this big circus to get rid of trading boredom (its my case)/

But almost all here (99.5%) are retail traders who will pump their $2000-$10000 accounts into strategies that they dont control neither master with leverages up to 500-800% with the hope of making about $2000-$5000 per month

Someone in the past told me that retail and pro are like keys for a lock bolt: one is the genuine key that will unlock that door, the other key is a fake one, very similar with the genuine one but it cannot unlock that door.

So, you dear Copernicus, with your sustainable risk-weighted returns....you are speaking for people from another planet, the Planet of PRO!

But....we all like this forum coz its fun and if you have a good liquour while you watch some threads, then you can really have some great moments
Attached Image
"What would you attempt to do, if you knew your success was a certainty?"
  • #1,623
  • Jul 12, 2017 4:17am Jul 12, 2017 4:17am
Quoting Fxentropy
{quote} But....we all like this forum coz its fun and if you have a good liquour while you watch some threads, then you can really have some great moments {image}
I'll drink to that :-)

Attached Image
  • #1,624
  • Jul 12, 2017 4:22am Jul 12, 2017 4:22am
Quoting Copernicus
{quote} I'll drink to that :-) {image}
Drinks are on me! ....running long pound...hope in a homerun like yesterday in dow

"What would you attempt to do, if you knew your success was a certainty?"
  • #1,625
  • Jul 12, 2017 4:36am Jul 12, 2017 4:36am
Dear C,

What is the rationale behind DTT's 24 hr and 1hr time frames? Did you test other time frames with this approach?

You said for EDTT factor 3 between time frames should be good.

Is it also true for DTTR and DTT?
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,626
  • Jul 12, 2017 5:03am Jul 12, 2017 5:03am
I have a feeling that somebody tricked me here:

Attached Image (click to enlarge)
Click to Enlarge Name: EexrdDp.png Size: 119 KB


Isn't it better to create some rule, how to evade these false breakouts?

For example here false breakout was 12 pips over last high which is 18% of the ADR.
So what about rule to put order 20% of ADR over last high/low?
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,627
  • Jul 12, 2017 5:27am Jul 12, 2017 5:27am
Quoting Copernicus
{quote} Yep R...the SDC in the DTTR is variable and has no defined value. We simply plot the SDC to fully encompass adverse historic volatility on D1. It therefore is a moving feast. EDIT. If you want to apply the principles below on H1 then I would recommend that you commence from D1 as a proxy for W1 before reverting to H1 for trade details. In a nutshell use D1 to define the broader trend conditions and then use H1 to define what is overbought and oversold in terms of the D1 definition. I haven't applied H1 with DTTR yet....but let logic be your...
DTTR 4h and 30m:

On your examples I could not see SDC less than 1.0. If we have very orderly trend as here on EURJPY:

Attached Image (click to enlarge)
Click to Enlarge Name: Yd3lXGR.png Size: 57 KB


Do we set SDC to less than 1.0 to touch lows by SDC? Or we only adjust it to values larger than 1.0?
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,628
  • Jul 12, 2017 6:03am Jul 12, 2017 6:03am
Quoting Ramadas
{quote} DTTR 4h and 30m: On your examples I could not see SDC less than 1.0. If we have very orderly trend as here on EURJPY: {image} Do we set SDC to less than 1.0 to touch lows by SDC? Or we only adjust it to values larger than 1.0?
Hi R. When we get to 2.5 x R then we adjust the SDC to 0.5 width. R is the distance between entry and initial stop. So 1.0 width until good profits at 2.5 x R then reduce by half. This ensures that SDC does not get too wide with more data and also does not leave too much profit on the table when trend ends. This applies to all variants except DTTR. :-)

PS. You are looking good on EURJPY.
  • #1,629
  • Jul 12, 2017 6:12am Jul 12, 2017 6:12am
Quoting Copernicus
{quote} Hi R. When we get to 2.5 x R then we adjust the SDC to 0.5 width. R is the distance between entry and initial stop. So 1.0 width until good profits at 2.5 x R then reduce by half. This ensures that SDC does not get too wide with more data and also does not leave too much profit on the table when trend ends. This applies to all variants except DTTR. :-) PS. You are looking good on EURJPY.
I meant before trade on DTTR, when you set up SDC on larger frame, you can have larger than 1.0 SDC like 1.7, 1.8. Can it happen that your SDC will be smaller than 1.0 if the trend is not volatile?
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,630
  • Jul 12, 2017 7:20am Jul 12, 2017 7:20am
Quoting Ramadas
{quote} I meant before trade on DTTR, when you set up SDC on larger frame, you can have larger than 1.0 SDC like 1.7, 1.8. Can it happen that your SDC will be smaller than 1.0 if the trend is not volatile?
Yep mate. However you need to apply a bit of discretion here. For a nice long term trend on DTTR you can find the SDC gets to say 0.5 or so with more data. That's ok and fine to trade. However for short term moves such as new trends those with a width of say 0.7 or so rarely last. You might want to give some more breathing room when the trend first starts out in a new direction. You might need to create your own rules here. I tend to stand aside from those and wait for more data.
  • #1,631
  • Jul 12, 2017 8:44am Jul 12, 2017 8:44am
Quoting Ramadas
I have a feeling that somebody tricked me here: {image} Isn't it better to create some rule, how to evade these false breakouts? For example here false breakout was 12 pips over last high which is 18% of the ADR. So what about rule to put order 20% of ADR over last high/low?
Hey R. I missed a few of your queries as I have been away this evening and posted by my phone. I am knackered mate and will respond tomorrow night if that's ok. Cheers C. May the trend be with you Obiwan :-)
  • #1,632
  • Jul 12, 2017 3:42pm Jul 12, 2017 3:42pm
Quoting j4d
A brief excerpt from a Kevin Davey Email..quite incredible.. " Quantopian is a backtesting platform, where users can upload strategies. So far, over 3,000,000 strategies have been uploaded. Of that large number, Quantopian has determined that only 8 are worth live trading! Wow!! My Strategy Factory process has a lot of fallout, but not that poor. I suspect many "professional" strategies were developed incorrectly..."
I reading here in the thread.

Game over for robots?
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,633
  • Jul 12, 2017 4:07pm Jul 12, 2017 4:07pm
Quoting Copernicus
Words of Wisdom from Peter Brandt - 50 + years of experience actively trading the markets Summarized transcript of major points: It took me 4-5 years before I became accustomed to the market and identify how to trade and build sufficient capital that I needed to trade. I don't know how many small accounts I blew up before I finally caught the scent for the approach to markets that suited my personality and made sense to me. I had to experience the blow-ups and make enough mistakes for...

Thank you for saving us time by writing all this.
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,634
  • Jul 12, 2017 4:12pm Jul 12, 2017 4:12pm
Quoting FXEZ
{quote} Hi Copernicus, I had to go back and find the definition of tail ratio because I wasn't familiar with that ratio. It makes sense - compare the most extreme 5% of profits to the most extreme 5% of losses. A system that has a high tail ratio seems more likely to have positive upside surprises while a system with a low tail ratio might have some nasty negative loss series that could materially affect the P&L of a system going forward. Maybe it is as you imply that when you look at the most extreme returns, the true nature of a system comes...
This is basically conclusion of Taleb's Fooled by Randomness....
Summary: http://jamesclear.com/book-summaries...-by-randomness
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,635
  • Jul 12, 2017 4:20pm Jul 12, 2017 4:20pm
Quoting FXEZ
{quote} Are 3 year track records meaningful? This is the blog post I referenced in the quote above. This post doesn't give a warm and fuzzy feeling about diversified programs that have Sharpe ratios < 0.5. {quote} After watching the video recommended by bassramy regarding historical drawdowns and their occurrence in a long-only buy and hold the market "strategy" with drawdowns existing about 75% of the time, while only 25% of the time new equity highs were made,...

So according to this blog post: https://blog.thinknewfound.com/2016/...ds-meaningful/
(above) quant should backtest 67 years of data to be sure, that the strategy is working?

Good luck with getting tick data...
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
  • #1,636
  • Jul 12, 2017 6:01pm Jul 12, 2017 6:01pm
Quoting Ramadas
Dear C, What is the rationale behind DTT's 24 hr and 1hr time frames? Did you test other time frames with this approach? You said for EDTT factor 3 between time frames should be good. Is it also true for DTTR and DTT?
Mate...Bottom line is that I had to start somewhere and my decision was made on the availability of long range data which was needed for testing purposes. There are so many potential variants of this approach mate and I suppose we have only really touched the surface. You will however get the picture though now that I am after an applied risk management logic to the strategies I select as opposed to a particular optimized setting relating to return performance. I avoid optimisation like the plague and if I use it, it is only for testing the very broad parameters of variables selected to ensure that variation around these broad variables do not produce a significant variation of returns. If they do, my hypothesis is that they are curve fit variables and I bin them.

Each of these variants have been developed according to an applied hypothesis dealing with risk management and they seem to stand up well in my testing. I am sure that other selected variables may outperform these tests historically....however that is not what I am after with a portfolio blend. The individual performance of each strategy, provided it has a positive edge over a broad range of market conditions is a minor consideration in overall performance. My attention is always focused on the risk management end as opposed to the return performance end. It is the blend in the portfolio that really matters and delivers the risk-weighted returns.
  • #1,637
  • Jul 12, 2017 6:07pm Jul 12, 2017 6:07pm
Quoting Ramadas
I have a feeling that somebody tricked me here: {image} Isn't it better to create some rule, how to evade these false breakouts?
Yep I had a look at my charts earlier and yesterday we experienced that damn mean reversion across these correlated markets......hasta la vista emerging trends. I just suck it in mate. Those false breakouts *** me off as well. I deal with them by placing my pending entries well away from current price to ensure that the only trends that interest me have a bit of history to back them up..... but in terms of my trend exits, I use the SDC like a ruler and just take the hit if the SDC is touched. My discretion is based around my entries only and not around my risk management principles which I stick to like glue.
  • #1,638
  • Jul 12, 2017 6:09pm Jul 12, 2017 6:09pm
Quoting Ramadas
{quote} I reading here in the thread. Game over for robots?
Never say die.....but it is a very tough game :-)
  • #1,639
  • Edited Jul 13, 2017 12:13am Jul 12, 2017 6:09pm | Edited Jul 13, 2017 12:13am
Quoting Ramadas
{quote} This is basically conclusion of Taleb's Fooled by Randomness.... Summary: http://jamesclear.com/book-summaries...-by-randomness
Yep mate. An excellent book which has inspired me a lot and the reason why you spend most of your time treading water if you have solid risk management techniques applied to simply deal with noise....with the occasional super profit on a non-random fat tail to make the wait and effort all worth it. :-)

PS That's an excellent review by the way :-)
  • #1,640
  • Jul 12, 2017 6:14pm Jul 12, 2017 6:14pm
Quoting Ramadas
{quote} So according to this blog post: https://blog.thinknewfound.com/2016/...ds-meaningful/ (above) quant should backtest 67 years of data to be sure, that the strategy is working? Good luck with getting tick data...

Hehehehe. :-)

The main take for me from this info is that no length of time or sample size will ever give you 100% conclusive proof (unless of course it is large enough to be representative of the entire population and not simply a sample)....but a larger sample size or a greater look back at least is better than a smaller sample size or shorter look back. If we avoid the maths of sample size for a moment.....the underlying logic is that a greater sample size or longer lookback deals with a greater variety of market conditions. I think it is the the array of different market conditions that is the important factor for testing robustness here as opposed to the actual numbers themselves.
5 traders viewing now