Strong May Results for Mechanical/Trend Following Systems

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 »

Congrats on the big year TK! Also, I commend you on posting your data. I am not in your camp of optimizing for recent performance. I prefer the approach of throwing a wide(r) net and running strategies similar to what Chelonia described above.
LeviF
Roundtable Knight
Roundtable Knight
Posts: 1436
Joined: Mon Dec 22, 2003 12:24 pm
Location: Des Moines, IA
Contact:

Post by LeviF »

TK wrote:On 1 October 2010...
What do you do with your existing positions at the end of each month?
rajivm
Roundtable Knight
Roundtable Knight
Posts: 131
Joined: Fri Jul 25, 2003 12:50 am
Location: New Delhi, India

Post by rajivm »

TK wrote:
These are: (1) betting strategy – while it is extremely difficult to improve upon the standard fixed-fractional betting strategy, it can be done provided you’re a private trader and don’t trade OPM – ironically, you have to have guts to do the reverse of what Eckhardt taught the Turtles....
Dear TK,
You need real guts to double up if in drawdown ( I think this is what you meant).....Have you been able to do so in real life?
Aaron01
Roundtable Knight
Roundtable Knight
Posts: 113
Joined: Thu Jun 02, 2011 1:21 pm

Post by Aaron01 »

rajivm wrote:
TK wrote:
These are: (1) betting strategy – while it is extremely difficult to improve upon the standard fixed-fractional betting strategy, it can be done provided you’re a private trader and don’t trade OPM – ironically, you have to have guts to do the reverse of what Eckhardt taught the Turtles....
Dear TK,
You need real guts to double up if in drawdown
I believe TK is saying he's taking off lots as the trades are moving in his favor instead of adding on more lots.
TK
Roundtable Knight
Roundtable Knight
Posts: 167
Joined: Tue Apr 15, 2003 5:45 pm

Post by TK »

LeviF wrote:
TK wrote:On 1 October 2010...
What do you do with your existing positions at the end of each month?
I am practical -- I keep them open and I exit them based on updated exit parameters.

Example 1
Let's say I trade a classic channel breakout system with 20-day entry and 10-day exit channels. I have an open position that is trailed with a 10-day channel stop. Then the month changes and a new lookback period for exits is 8 days. I keep my position open but from now on my trailing stop is the 8-day channel.

Example 2
Let's say I use time-based exits. All open trades are exited after 10 days. I have an open trade that is 8 days old at the month end. After re-optimization it turns out that the new value for my time-based exits is 5 days. I exit my trade at the next open because it is 8 days old which is longer than the current exit value.

I find that approach practical and easy to implement. I know that the resulting equity curve is not exactly equivalent to what you theoretically get when you stitch chunks of individual equity curves such as in the built-in walk-forward code in TBB. But rebuilding your entire portfolio every month would kill you in terms of transaction and slippage costs if you traded short and medium timeframes. So I like to keep it simple and practical and it works for me.
TK
Roundtable Knight
Roundtable Knight
Posts: 167
Joined: Tue Apr 15, 2003 5:45 pm

Post by TK »

rajivm wrote:
TK wrote:
These are: (1) betting strategy – while it is extremely difficult to improve upon the standard fixed-fractional betting strategy, it can be done provided you’re a private trader and don’t trade OPM – ironically, you have to have guts to do the reverse of what Eckhardt taught the Turtles....
Dear TK,
You need real guts to double up if in drawdown ( I think this is what you meant).....Have you been able to do so in real life?
Years back, before the TBB forum was launched, the No. 1 forum for mechanical traders was Chuck Lebeau’s TraderClub. Lots of good stuff was posted there over the years but today, sadly, that forum is gone. Before it was closed, however, I had saved a couple of good threads to my disk for future reference. The post below, on money management by Leonardo, is one of many I had managed to save:
On 13 Sept., 2001, on the TraderClub forum, Leonardo wrote: My money management methods are fairly straightforward. After I have determined the edge of my system; throwing all the data I have at it; 20 to 40 years on over 40 markets-- good trading markets and bad--- I trade a fixed fraction of capital at about 1/3 of the Kelly Criterion number, (otherwise known as optimal f) and let her buck. Even though this is fairly conservative, it is still possible to get 40% and 50% drawdowns on the rare occasion. Don't like it, but I understand that the markets must come at you if you are trading correctly. The mark of a GOOD system and MM combination is..., being able to pull out of the drawdown and go to new equity highs.

I will also mention a discussion I had with a friend/trader who has a small fund underway. We were talking about optimal forms of money management, and how difficult it really is to improve on a fixed fraction money application. The problem is: the inability to tell in advance the relative edge you are facing on this next trade versus the edge you had on the last one or the trade you will take in the future. And then I told him how some of the ways private traders trade have certain advantages over the way a fund has to trade.

In games like blackjack or even roulette, the odds-per-money-decision change quite dramatically, and are possible to be identified in advance (with great skill and special circumstances) because of changes in the conditions of play.

The odds change from trade to trade also, but it is not as easy to differentiate the edge of one trade over another due to the process of finding a winning trading method. Also, not so many runs are available in fresh data to have a statistical significance in one subset of trades over another.
Now think about it. In blackjack, if you want to maximize your gains or even risk-adjusted gains, you have to vary you bets percentage-wise based on the current odds. It is fairly easy in blackjack because you can estimate/calculate the current odds for every possible hand. It is much harder in trading where the odds of the very next trade is pretty much unknown. If you could know that the odds of this trade was better than the odds of that trade, you could vary your percent risk per trade accordingly.

I asked Leonardo about it one day and he confirmed that all great private traders trade simple systems that have the greatest edge when in a drawdown.

The good news is that it is testable and the idea is very easy to test in TBB when you start experimenting with negative values for Drawdown Reduction Amount. Let’s say that you normally risk 2% per trade. Try to code up the betsize formula by which you will risk only 1% per trade when your drawdown is 0% (new equity high) and as much as 3% per trade when in a 40%+ drawdown. For most systems that I have tested and that I trade in real life, increasing your percent risk per trade in a DD will improve you risk-adjusted return and shorten your Longest DD.

And yes, I trade that way in real life.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

I have a large collection of posts both from Chuck Le Beau's old site and from the old Trading Recipes list. I have not looked at either collection for some years. It would be interesting to see how some of the other posts have stood up to the test of time.
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B »

AFJ Garner wrote:I have a large collection of posts both from Chuck Le Beau's old site and from the old Trading Recipes list. I have not looked at either collection for some years. It would be interesting to see how some of the other posts have stood up to the test of time.
I remember way back ~1995 when Chuck was thinking about starting that site and asking about my thoughts, etc. Time flies. Sadly, I recall that site admin wasn't taking seriously enough and some sort of virus/scammers/etc took over the site leading to its demise? I think that's probably been 10 years now too. Prior to that it was mostly Club3000 and CompuServe boards...probably pre-dates many on here. Having 9600 baud modems hit the scene sure allowed much quicker downloads from CompuServe (I used some really nice program that managed download/uploads to CompuServe forums).
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B »

Back on topic at hand -- as discussed elsewhere on this board recently (perhaps in the section that is now banned from those of us not owning TB...even if we contributed significantly to the original s/w design), I do think it is a bit of a stretch to even begin to compare the history of market data to the current environment that exits. It's easy to take the party line and say nothing has changed, I'm going to trade as before, etc. Back in the mid-90s when I was deep into long term TF design, that is exactly the angle I'd argue with anyone. Now however, times actually *are* different. Why?

Well, if you look at total credit outstanding (public + private), specifically the slope of that curve for the past 40 years, it can be fit to an exponential equation. Credit was not only on a continual expansion but was increasing at an increasing rate for decades. It reached a crescendo in 2007, and the time since then has seen central banks and governments trying to prop up the corpse of credit growth to "get back to where we were" for whatever good that will actually do.

Hence the actual market environment that has existed for decades, credit growth + credit rate of change growth on a continual basis, is now derailed. The markets are reflecting fake pricing based on money printing and interventions/manipulations. This fake pricing has/will go on long enough to convince the masses that it is normal.

Where does all this lead? Who knows...
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

Chuck B wrote:Back on topic at hand -- as discussed elsewhere on this board recently (perhaps in the section that is now banned from those of us not owning TB...even if we contributed significantly to the original s/w design)..
Hmm, I don't enjoy censorship either but then I must realize that I have no ownership rights over this site and can be banned from posting at any stage. Rather like Facebook, the users contribute much to the site and the profits of its owners but do not have any ownership rights themselves. I can not of course blame the proprietors - we all need to make a living and at times I must admit that I have gone overboard on criticism of one particular fake Guru who exudes snake-oil.

Back to the point in hand. I guess you are right about the past 40 years but I do wonder whether that invalidates LTTF. I have often lamented the lack of data over the past 6 months and perhaps to obtain such data as may be available for the past 200 years might prove worthwhile, even if that data is monthly at best.

I know that much of what we see happening at present seems controlled and manipulated and I guess it must be true that man has never lived on credit alone, as he seems to have done recently, but there again, history never seems to reveal anything much that is new.

Free trade itself is a fairly new phenomenon. I remember studying the Repeal of the Corn Laws. I think I may be right in saying that regulation and control has been the rule rather than the exception in human history. The gold standard, pegged exchange rates, import controls, the inability to take currency out of your home country, 90% tax rates at times. Periods of staggering inflation, periods of equally staggering deflation such as when Europe was devastated by the Black Death.

I do think, believe, judge back testing to be largely a mirage, especially on only 40 years of data. My own personal experience has been one of continuous self deception. I still however have a residual belief in TF. I still believe that reasonable leverage can probably keep you in the game through the bad times and I also believe you can deal with the wicks and wild intra-day movements in various ways if you are a long term trader.

What I am guilty of, is ignoring both my own advice (I wince when I look back at some of my more sensible posts over the years) and the psychological aspects which I have always rather scorned (and wrongly so). I have been beguiled by greed time and time again and then gripped by fear; I have done things I ought not to have done and left undone those things I ought to have done. (I do love the Book of Common Prayer!).

I do think that TF is highly unstable; I do believe (and it has been my unfortunate experience) that drawdowns will always be substantially greater than shown in that beguiling back test. I also believe however that if you can stand the heat you will probably win through. Sure, the ride will be rougher than in the halcyon days when hedge funds were unheard of but I still believe it works over time. All that hoary old crap is correct; don't trade too big nor with too much of your capital, live within your psychological limits, live to fight another day, fear will inevitably follow greed if you let the latter get out of hand etc etc.

Who knows, as you say. Not I, that is for sure. But I intend to struggle on in a hopefully wiser and more measured manner.
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B »

I should also mention the other possibility in mind -- the possibility that long term TF works better than it ever has during this continual fake pricing environment. :D What better action for markets to take other than to shake out all the doubters and easy drifters right before a period of performance unlike any seen in the past occurs? :lol: :shock: :wink: You know, just as all the "professionals" have finished tailoring their methods to "adapt" to the current "environment." :P
akhoury
Roundtable Knight
Roundtable Knight
Posts: 113
Joined: Mon Feb 28, 2011 11:07 am
Location: Stuart, Florida

Dinner Contest on Original Thesis

Post by akhoury »

This thread has covered so many good points and seriously arguable philosophies on markets, intervention, and data change…One important addition to the philosophical argumentation that I have not seen is that the fiat era is totally dependent on central management and intervention (so we need to get over it!)…rather than debate philosophy I was really trying to get a quantitative, empirical argument –both pro and con—for the thesis that short term systems are better post Lehman. I lack some of the technical skills that many contributors have, but as a group we have an ability (some might say edge) to rapidly exchange quantitative argumentation and through the process educate each other.

So, I am going to issue a challenge and offer a contest: I will buy dinner for 2, at the restaurant of the winner’s choice, for the best quantitative argument, pro or con, the thesis above. Here are some rules:

1. Only one point of view and two posts allowed per entrant.
2. The second post can only be a rebuttal in support of your argument.
3. I will ask Tim and Jake to help with the judging (unless they would prefer to compete for the dinner)
4. Wine capped at a total of $300..
5. Contest ends this week by Friday 5:00pm est
6. Winner declared within the week thereafter
7. I fully expect some of our more advanced players to think this is stupid—that’s ok—somebody, maybe a junior member, is going to win.
8. Philosophical expressions—beyond your thesis -- will deduct from your score…
9. Submissions and questions will only be accepted via the forum so it is fair to all who want to play…

Good Luck and have Fun…ack
Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday »

Chuck,

You very eloquently said what I was trying to say in a previous post on this topic. IMHO, the past 10 to 30 years of data will tell you that a long strategy has risk adjusted returns far in excess of a short strategy when using LTTF. Moving forward I am spending most of my time improving my short strategies as I feel that de-leveraging will move the short and long risk adjusted returns closer together. I know that this comes across as a prediction but there are several times when trend following that you run into inconsistencies (ie the world is not black and white).

TK,

Thanks for sharing your data. I can say that your experience is inspiring, in that someone has the fortitude to live through the performance you presented. I believe fortitude is one of the most desirable characteristics of a profitable trader. And, I think, pushing the bounds of trend following in terms of position sizing and parameter selection is a good thing. Best of luck to you.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

Toosday wrote: IMHO, the past 10 to 30 years of data will tell you that a long strategy has risk adjusted returns far in excess of a short strategy when using LTTF. Moving forward I am spending most of my time improving my short strategies as I feel that de-leveraging will move the short and long risk adjusted returns closer together.
One area which makes combined long and short trading look worse (sometimes far worse) over the past decade is where you use risk controls/limits. I often find that an unconstrained system (in terms of risk) trades better in back testing using L/S than long only.

Trading short has been far less profitable than trading long and thus where you have risk controls in place less profitable short trades will take risk availability from more profitable long trades.

Moving on from that it is interesting to control risk by the size and composition of the portfolio (rather than refusing trades on a large portfolio when risk availability has been used up).

I am not suggesting curve fitting. Use the random portfolio generator and assign (by way of example) 5 contracts to each sector (IE 5 softs, 5 grains etc)...or 4 or 6 etc. It is one possible argument against the jumbo portfolio and for a move back towards smaller portfolios of 30 or 50 different contracts.
marriot
Roundtable Knight
Roundtable Knight
Posts: 365
Joined: Thu Nov 20, 2008 3:02 am

Post by marriot »

But if you have all the possibile instrument in portfolio, for sure you are not cheathing.
Last edited by marriot on Thu Jun 14, 2012 1:49 pm, edited 2 times in total.
marriot
Roundtable Knight
Roundtable Knight
Posts: 365
Joined: Thu Nov 20, 2008 3:02 am

Post by marriot »

Whatever system you test, starting from 2000, DD begin to be bigger.
For what is worth, i think that changes are related to Internet.
Trading made easy for everyone plus automated trades for the most skilled people.
Last edited by marriot on Thu Jun 14, 2012 1:43 pm, edited 1 time in total.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

marriot wrote:Whatever system you test, starting from 2000, DD begin to be bigger.
Yes, I think markets have changed for all the reasons so often discussed. But I would raise one point I am sure you must be aware of: very few markets existed back in 1965 and more new markets have come on-stream right into the past decade. If, therefore, you are running a system from 1965 to date with static parameters using a portfolio of (say) 190 instruments, it is hardly surprising if max DD increases as more instruments come on stream. Did return increase in line with drawdown, or did the risk return ratio deteriorate - that perhaps is the important question.
marriot
Roundtable Knight
Roundtable Knight
Posts: 365
Joined: Thu Nov 20, 2008 3:02 am

Post by marriot »

We do not have Data from preihstoric men sending messages with smoke.
But with what we have, we can find best parameters from 1965 to 2008, from 1990 to 2008, from 2000 to 2008 and finally see what set is robust.
It is my opinion that "More" is not always better.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

All things being equal, trading more instruments over time (as they come on-stream) with a static system and without the use of risk limitation is likely to increase the maximum draw down at the end of the run. That may be one reason why your draw down chart looks the way it does. Not the only reason, for sure.....but..........

I don't think you need to go back quite as far as the stone age to make the point but if you happen to have the data to hand for corn or wheat from 5 or 10,000 BC I'd be grateful if you would put it in a handy size zip file and email it to me.
rabidric
Roundtable Knight
Roundtable Knight
Posts: 243
Joined: Mon Jan 09, 2006 7:45 am

Post by rabidric »

marriot wrote: It is my opinion that "More" is not always better.
Less is more? Actually, I strongly agree. If you know how there are quite a few ways to skin a lesser cat to achieve greater yield, some markedly better than others though. Using a modern lawnmower indiscriminately may leave a nasty mess with neither skin or meat salvageable. But using some stone age flint knife by hand in the appropriate fashion may yield surprisingly clean results. The real question is, do you have the stomach for catmeat? I personally prefer it in the short-term.

bring back the stone age?! :? :wink: :lol:
Post Reply