The Mirage

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
trackstar
Roundtable Fellow
Roundtable Fellow
Posts: 76
Joined: Sun Mar 09, 2008 11:03 pm
Location: Houston Texas

Post by trackstar » Thu Jun 14, 2012 9:43 am

What timeframe were you trading? Daily or weekly?

Aaron01
Roundtable Knight
Roundtable Knight
Posts: 113
Joined: Thu Jun 02, 2011 1:21 pm

Post by Aaron01 » Fri Jun 15, 2012 1:02 am

Imperium wrote:Here is the link to the article mentioned:

http://www.canada.com/mobile/iphone/tec ... story.html

It was barley and sesame prices, not wheat.

This is what you're up against.
Up against what? We have no idea what purpose the data was used for and whether or not it materially impacted Winton's returns; i.e. whether the benefit obtained was sufficient with the effort of the exercise.

I would hazard a guess that such a task probably serves its greatest role as a marketing point. Much of the data may have been useless for statistical testing purposes, but to some less informed (or less cynical?) capital allocators it may be an indication that Winton leaves "no stone unturned."

Perhaps everything is not as it seems....

squaredQ
Full Member
Full Member
Posts: 23
Joined: Wed Feb 01, 2012 5:27 pm

Post by squaredQ » Fri Jun 15, 2012 3:19 am

Thanks for sharing some specifics.

I wonder if in hindsight whether or not you could go backwards and account/isolate the factors that dramatically changed the characteristics over the out of sample period.

I would hazard a guess that correlation dramatically and unexpectedly synced up at the onset of the precipice, possibly giving an early warning signal? There was a great animation study done on that phenomenon some time back (although more directly related to the 08 meltdown).

Imperium
Roundtable Fellow
Roundtable Fellow
Posts: 52
Joined: Tue May 17, 2005 1:04 pm
Location: London

Post by Imperium » Fri Jun 15, 2012 5:23 am

Aaron,

I too believe it is some 'no stone unturned' marketing spiel. But it also highlights the fact that people are looking into all sorts of data to try to get an edge from somewhere. Markets are jumpy and nervous, they have been for a few years and its likely this will continue for the next few years. Old trading certainties are like a square peg being jammed into a round hole.

I think some people are unable to understand that the current macro environment is very different to ones we've seen in the past - by 'past' i mean the 30 years of daily data everyone is using to build their trading strategies.

Trend following isn't broken, it hasn't come to an end and some recent posts on this forum have very clearly pointed out whats changed. Understand and adapt or die.

If you are a LT trader then you have to understand the effects of ZIRP, CB interventions, potential economic chaos/collapse in the richest trading block in the world (the EU), infantile political posturing in the US (and EU), unprecedented global deleveraging, etc.

Each of these is a significant event, we're experiencing them all.

illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

Post by illuminati » Fri Jun 15, 2012 12:52 pm

IMO....TF is becoming a overcrowded arena. The number of CTAs, the number of individual traders, all have made the business and strategy harder. TF can easily be replicated and due to this low entry barrier, everyone is running a similar strategy.

I think to make money onwards in TF, ether you reinvent the wheel somehow and apply it in a novel way or just through your money at some CTA of some size and reputation. Or else, those 50%+ DD are unavoidable when everyone is using MAs and Breakouts.

But...humans are confident creatures. Each new trader or existing ones will convince themselves their "system" is special and different. And this is where trouble emerges when everyone think they are "different" when all in all, your doing the same thing like everyone else.

Trading Leech
Senior Member
Senior Member
Posts: 48
Joined: Mon Nov 29, 2010 2:07 pm

Post by Trading Leech » Fri Jun 15, 2012 6:09 pm

(Much of what I'm posting below is a variation of what has already been said here and in other threads.)

A bit off-topic:

I think people need to let go of the idea that trend following is not prediction. It is prediction, just like any other trading strategy out there (including the multitude of successful strategies mentioned in Schwager's Market Wizards books). It is, indeed, predicting that the trends will continue in a similar way as they have done in the past.

It's too bad that the "gurus" don't mention this, which is understandable given that they make their money not by trading but by convincing people that trend following is some kind of a "holy grail" when in fact it is not. (This is not to say that I am not planning on doing trend following in the future with some of my capital.)

Getting back to this thread:

Yeah, this is getting quite a dangerous game indeed. More and more people are getting into the same game. I'm not into it yet. I'm still doing all the research I believe is necessary to carry out before I start. I think it will take me at least one more year. And to tell the truth, I'm getting more and more tense as time goes by.

I'm also becoming more and more skeptical of historical data. It only provides a very rough idea of what the future may look like. I like what rhc and others have to say about trying to create future scenarios. A part of this is generating one's own data using, say, "data scrambling" as discussed in chapter 8 of Chande's "Beyond Technical Analysis". (Thanks to rhc for suggesting this here.) Finally, if one is still not sure, doing further overestimation of losses and underestimation of profits will provide even more safety. The problem is, of course, that all of this safety may end up bringing one down to a way too low a position size, thereby eliminating future returns.

I no longer have any illusions about how difficult and risky this game is going to be. In fact, the more I imagine my first trade, the more frightened I get :?

illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

Post by illuminati » Fri Jun 15, 2012 11:34 pm

Trading Leech wrote:(Much of what I'm posting below is a variation of what has already been said here and in other threads.)

A bit off-topic:

I think people need to let go of the idea that trend following is not prediction. It is prediction, just like any other trading strategy out there (including the multitude of successful strategies mentioned in Schwager's Market Wizards books). It is, indeed, predicting that the trends will continue in a similar way as they have done in the past.

It's too bad that the "gurus" don't mention this, which is understandable given that they make their money not by trading but by convincing people that trend following is some kind of a "holy grail" when in fact it is not. (This is not to say that I am not planning on doing trend following in the future with some of my capital.)
Pot"a"to ...pot"ah"to Its all about convincing yourself where you stand.

I cant speak for others but personally and in my own opinion, there is high chance of failure in individual TF due to the fact that successful TF cant be done with < $5 million. that is if you want to diversify across time frame, markets, and systems traded. Individual traders going to trade one TF system on a basket of markets don't have a wide ""margin of safety" as the probability of your system suffering from DD or some unpredictable behaviour increase. Jez Liberty does a trend following report on his site and if you look at any one month, each system varies widely in performance.

Limited capital is the foremost reason why TF is hard for individuals traders. Even if you got 5 million laying around, you better have some novel ideas. (Thorp = combining fundamental and technical)

alp
Roundtable Knight
Roundtable Knight
Posts: 309
Joined: Mon Aug 27, 2007 8:09 pm

Post by alp » Sun Jun 17, 2012 12:16 am

Imperium wrote:Trend following isn't broken, it hasn't come to an end and some recent posts on this forum have very clearly pointed out whats changed. Understand and adapt or die.
Or perhaps we're in the process of the markets getting rid off most trend followers until TF starts to work again? What kind of adaptation do you refer to?

marriot
Roundtable Knight
Roundtable Knight
Posts: 347
Joined: Thu Nov 20, 2008 3:02 am

Post by marriot » Sun Jun 17, 2012 1:59 am

Tb as it come.
All systems as they come, no parameters changes.
160 different futures, no Lme.
Basic money manager blox with 2 fixed contracts.
Draw your own conclusion.
Attachments
2 contratti.JPG
2 contratti.JPG (129.89 KiB) Viewed 5380 times

Aaron01
Roundtable Knight
Roundtable Knight
Posts: 113
Joined: Thu Jun 02, 2011 1:21 pm

Post by Aaron01 » Sun Jun 17, 2012 11:10 pm

alp wrote:
Imperium wrote:Trend following isn't broken, it hasn't come to an end and some recent posts on this forum have very clearly pointed out whats changed. Understand and adapt or die.
Or perhaps we're in the process of the markets getting rid off most trend followers until TF starts to work again? What kind of adaptation do you refer to?
Markets are probably no different than environments. If a species doesn't adapt to a changing environment it will die off. That may mean using a different method to hunt, or just hunting for different prey...

alp
Roundtable Knight
Roundtable Knight
Posts: 309
Joined: Mon Aug 27, 2007 8:09 pm

Re: The Mirage

Post by alp » Mon Jun 18, 2012 11:59 am

AFJ Garner wrote:Yes, the system worked well for over 30 years and then clapped out the day I started trading. How was I to know? With hindsight, what should I have done? What should I have learnt from the exercise?
With hindsight everything is as clear as crystal. You just change the system and get different backtest results, and have your system working again.

I attach pictures of the equity curve of my old system, which I abandoned 26 months after running it with a small capital, as a real life test.

What I should have learnt is up to me. But then I wonder how much lifetime I have got left to keep learning it and implementing a new system along the way. Perhaps you've got this answer for me? :)
Attachments
Old System Backtest - 19800102 to 20041231.jpg
Old System Backtest - 19800102 to 20041231.jpg (92.44 KiB) Viewed 5252 times
Old System - 20050102 to 20120615.jpg
Old System - 20050102 to 20120615.jpg (101.9 KiB) Viewed 5252 times

Aaron01
Roundtable Knight
Roundtable Knight
Posts: 113
Joined: Thu Jun 02, 2011 1:21 pm

Re: The Mirage

Post by Aaron01 » Mon Jun 18, 2012 3:44 pm

alp wrote:
AFJ Garner wrote:Yes, the system worked well for over 30 years and then clapped out the day I started trading. How was I to know? With hindsight, what should I have done? What should I have learnt from the exercise?
With hindsight everything is as clear as crystal. You just change the system and get different backtest results, and have your system working again.

I attach pictures of the equity curve of my old system, which I abandoned 26 months after running it with a small capital, as a real life test.

What I should have learnt is up to me. But then I wonder how much lifetime I have got left to keep learning it and implementing a new system along the way. Perhaps you've got this answer for me? :)
Do you have the 2nd chart scaled logarithmically?

alp
Roundtable Knight
Roundtable Knight
Posts: 309
Joined: Mon Aug 27, 2007 8:09 pm

Re: The Mirage

Post by alp » Mon Jun 18, 2012 4:17 pm

Aaron01 wrote:Do you have the 2nd chart scaled logarithmically?
As the thread's title is The Mirage and perhaps this is a pretty good example of a big mirage, I include equity curves in both scales.
Attachments
Old System - 19800102 to 20120615 - Log Scale.jpg
Old System - 19800102 to 20120615 - Log Scale.jpg (92.74 KiB) Viewed 5199 times
Old System - 19800102 to 20120615 - Linear Scale.jpg
Old System - 19800102 to 20120615 - Linear Scale.jpg (117.81 KiB) Viewed 5199 times

stopsareforwimps
Roundtable Knight
Roundtable Knight
Posts: 199
Joined: Sun Oct 10, 2010 1:47 am
Location: Melbourne Australia

Post by stopsareforwimps » Wed Jun 20, 2012 10:17 pm

alp wrote:
Imperium wrote:Trend following isn't broken, it hasn't come to an end and some recent posts on this forum have very clearly pointed out whats changed. Understand and adapt or die.
Or perhaps we're in the process of the markets getting rid off most trend followers until TF starts to work again? What kind of adaptation do you refer to?
This looks interesting.

http://www.columbia.edu/~kd2371/papers/ ... d/mom4.pdf

If trend following is like buying out of the money options, then such a strategy would be unprofitable when options are expensive.

In the context of TF expensive options would equate to highly volatile returns.

stopsareforwimps
Roundtable Knight
Roundtable Knight
Posts: 199
Joined: Sun Oct 10, 2010 1:47 am
Location: Melbourne Australia

Re: The Mirage

Post by stopsareforwimps » Fri Jul 06, 2012 7:43 pm

AFJ Garner wrote:In an effort to gaze into the crystal ball ...
Another issue is the veracity of prices used to settle futures contracts.

http://www.nakedcapitalism.com/2012/07/ ... tives.html

Bravochico
Full Member
Full Member
Posts: 24
Joined: Wed Sep 14, 2011 11:46 pm
Location: Southern California, Ca

Post by Bravochico » Sun Jul 08, 2012 12:45 pm

Just some observations

(typing on iPad watching Fed Murray, so excuse typos)

There's no mention of exit issue with trend following. The trends are still here, but the "give backs" in unrealized pnl much more violent than before.

Sure there are more false breakouts, but IMHO the real damage to lttf is the increased veracity of the givebacks.

Also, this whole argumentof buying cheap options being only way to make money is much too general to be of practical use. I've seen many options be cheap ad stay cheap. Vice versa.

I haven't read the paper yet, but at surface it appears a thesis with many conditional footnotes. Most academics,researchers don't excel in trading. This business is half art. That's part they sometimes miss.

stopsareforwimps
Roundtable Knight
Roundtable Knight
Posts: 199
Joined: Sun Oct 10, 2010 1:47 am
Location: Melbourne Australia

Post by stopsareforwimps » Sun Jul 08, 2012 9:36 pm

stopsareforwimps wrote:
alp wrote:If trend following is like buying out of the money options, then such a strategy would be unprofitable when options are expensive.

In the context of TF expensive options would equate to highly volatile returns.
I've been rereading the Market Wizards books. It was striking how many traders do something in the vein of cutting size after painful losses, or taking volatility holidays, or they have daily/weekly/monthly loss limits that cause a cessation of trading.

Usually the rationale is that you cannot be rational after big losses and you will just make mistakes.

There may be another reason. This paper "Momentum Crashes" by Kent Daniel and Tobias Moskowitz

http://papers.ssrn.com/sol3/papers.cfm? ... id=1914673

argues that after big downdrafts in markets, momentum investing is predictably unprofitable.

A big downdraft is going to have a trend in it so trend followers are likely to be short at the end. The trend is usually followed by a painful reversal, or by listless trading, which will also create losses. Either way, going small until you are profitable again would make sense.

stopsareforwimps
Roundtable Knight
Roundtable Knight
Posts: 199
Joined: Sun Oct 10, 2010 1:47 am
Location: Melbourne Australia

Post by stopsareforwimps » Tue Jul 10, 2012 8:55 pm

stopsareforwimps wrote:This paper "Momentum Crashes" by Kent Daniel and Tobias Moskowitz

http://papers.ssrn.com/sol3/papers.cfm? ... id=1914673

argues that after big downdrafts in markets, momentum investing is predictably unprofitable.
Similar paper here

Tail Risk in Momentum Strategy Returns

Abstract:
Price momentum strategies have historically generated high positive returns with little systematic risk. However, these strategies also experience infrequent but severe losses. During 13 of the 978 months in our 1929-2010 sample, losses to a US-equity momentum strategy exceed 20 percent per month. We demonstrate that a hidden Markov model in which the market moves between latent "turbulent'' and "calm'' states in a systematic stochastic manner captures these high-loss episodes. The turbulent state is infrequent in our sample: the probability that the hidden state is turbulent is greater than one-half in only 20% of the months in our sample. Yet in each of the 13 severe loss months, the ex-ante probability that the hidden state is turbulent exceeds 70 percent. This strong forecastability accentuates the price momentum puzzle.

http://papers.ssrn.com/sol3/papers.cfm? ... id=2076622

Post Reply