Will you survive?
-
- Senior Member
- Posts: 38
- Joined: Wed May 28, 2003 2:17 am
Will you survive?
1994 is the worst period for long term trend following in my testing. Stock markets were consolidating, interest rates were reversing, and on average things were flat for the rest. Too much leverage was fatal in this period, and funds that experienced losses saw a client exodus like they'd never seen before.
A casual browsing of the IASG database indicates that trend following funds are once again flush with new money as a result of the great trends of the last 10 years. The bubble and subsequent bust in stocks and a more recent cyclical commodities boom marked by huge bull markets in energy, agriculture, and even stocks have created great environments for trend followers.
But I anticipate the next global lull in price changes could be coming in the next few years because the commodities bull is eventually going to plateau. Markets like copper, crude oil, and wheat are going to stay high, but at this point they are vulnerable to consolidation and one day some fundamentals maverick is going to come out with all the reasons they are overpriced. Most commodities markets have responded to the upside at this point, and it's this idea that all real assets are better than equities that can create fatally high correlations.
The dollar has also been punished enough that extended periods of sideways action is conceivable.
Stock markets around the world are overvalued (even the Shanghai Composite has a PE ratio of 60). If they don't go down they will be flat.
Interest rates will go higher as the US and Europe battle for capital in an inflationary world that is looking to Asia for growth. A sustained downtrend in bonds could create volatility in the currency markets (a good thing for trend followers).
Of course, predictions of these events are outside a purist trend follower's scope. The point isn't predictions but rather that an unfriendly, supposedly low-probability environment such as 1994 will come as a surprise after a long period of success. It might happen this year or not for another 5 years, but at some point the world is going to be in-between trends.
If a strategy is only as good as the worst case scenario and the biggest drawdown is always around the corner it's important to think about why a trend follower might be able to survive periods of global trend reversals or low volatility.
1. Consider that too much leverage can single handedly crush the chances of a strategy working out. LTCM references are overdone, but consider that John Meriwether founded a new fund that is using the old LTCM arbitrage strategies. The only difference is the amount of leverage. "Whatever you think your position ought to be, cut it in half." -- Bruce Kovner.
2. One of the unheralded keys to survival and performance in volatility systems is to diversify by trading more uncorrelated markets. This is the reason Transtrend is beating everyone else's pain/gain ratios. They are trading everything under the sun.
A casual browsing of the IASG database indicates that trend following funds are once again flush with new money as a result of the great trends of the last 10 years. The bubble and subsequent bust in stocks and a more recent cyclical commodities boom marked by huge bull markets in energy, agriculture, and even stocks have created great environments for trend followers.
But I anticipate the next global lull in price changes could be coming in the next few years because the commodities bull is eventually going to plateau. Markets like copper, crude oil, and wheat are going to stay high, but at this point they are vulnerable to consolidation and one day some fundamentals maverick is going to come out with all the reasons they are overpriced. Most commodities markets have responded to the upside at this point, and it's this idea that all real assets are better than equities that can create fatally high correlations.
The dollar has also been punished enough that extended periods of sideways action is conceivable.
Stock markets around the world are overvalued (even the Shanghai Composite has a PE ratio of 60). If they don't go down they will be flat.
Interest rates will go higher as the US and Europe battle for capital in an inflationary world that is looking to Asia for growth. A sustained downtrend in bonds could create volatility in the currency markets (a good thing for trend followers).
Of course, predictions of these events are outside a purist trend follower's scope. The point isn't predictions but rather that an unfriendly, supposedly low-probability environment such as 1994 will come as a surprise after a long period of success. It might happen this year or not for another 5 years, but at some point the world is going to be in-between trends.
If a strategy is only as good as the worst case scenario and the biggest drawdown is always around the corner it's important to think about why a trend follower might be able to survive periods of global trend reversals or low volatility.
1. Consider that too much leverage can single handedly crush the chances of a strategy working out. LTCM references are overdone, but consider that John Meriwether founded a new fund that is using the old LTCM arbitrage strategies. The only difference is the amount of leverage. "Whatever you think your position ought to be, cut it in half." -- Bruce Kovner.
2. One of the unheralded keys to survival and performance in volatility systems is to diversify by trading more uncorrelated markets. This is the reason Transtrend is beating everyone else's pain/gain ratios. They are trading everything under the sun.
Last edited by Eric Winchell on Wed Jun 11, 2008 7:58 pm, edited 2 times in total.
Just yesterday I was looking at some simulations that go back 35 years, which may present another perspective on "the worst period" for systematic trading. These simulations tested four different parameter settings for a single underlying system; they were exploring different choices for "speed" (frequency of trading), "greed" (letting your profits run vs. tightening the stops), and so forth. The Blox calculated Annual Returns in % were collected into a single graph. It turns out to be fairly easy to generate this kind of plot with Blox; you just export the "Yearly Performance Summary" to a spreadsheet, and then create a multi-column plot in the spreadsheet.
As the chart shows, this particular set of systems suffered their worst year in 1980. March 1980 to be precise. That was when Nelson Bunker Hunt's corner of the Silver market collapsed and Silver went limit-down for something like 13 days in a row. The other metals suffered similar fates. Luckily, this collapse occurred after a huge run-up, (Hunt's corner was "working" well, for a while) so the system was "merely" giving back profits it just collected on its metals positions, three months before. The systems were run on a "trade them all" portfolio of about 100 global markets, but the system rules include total-risk limiters and sector-risk limiters that aim to make the systems tradeable for a $200K account size.
For these systems, the year 1994 was relatively unremarkable. Neither better nor worse than average. For what it's worth.
As the chart shows, this particular set of systems suffered their worst year in 1980. March 1980 to be precise. That was when Nelson Bunker Hunt's corner of the Silver market collapsed and Silver went limit-down for something like 13 days in a row. The other metals suffered similar fates. Luckily, this collapse occurred after a huge run-up, (Hunt's corner was "working" well, for a while) so the system was "merely" giving back profits it just collected on its metals positions, three months before. The systems were run on a "trade them all" portfolio of about 100 global markets, but the system rules include total-risk limiters and sector-risk limiters that aim to make the systems tradeable for a $200K account size.
For these systems, the year 1994 was relatively unremarkable. Neither better nor worse than average. For what it's worth.
- Attachments
-
- Yearly returns for 4 different systems
- wildcat.png (11.92 KiB) Viewed 25833 times
-
- Sending the Yearly Performance Summary data to a spreadsheeet
- exxportt.png (38.44 KiB) Viewed 25835 times
Cograts on being able to predict the future ?
Another scenario you may wish to consider
Trend following has actually been awful / ish for the last 10 years ? Look at the performances of funds like Dunn and JWH over the last 5 years compared to their early years ?>
Perhaps we are on the verge of a really good trend following environment where trend followers make 100-200 % annual returns ? Why not > The whole world is awash with money in short term systems that try to capture small moves and return 10 % a year
From my calculations the hedge fund industry currently manages 2.9/3.2 trillion dollars with trend followers accounting for less than 2.4 % of that - the real risk is proper volatility for the next 10 years that teaches investors a lesson they badly need to learn - There aint no such thing as 10 % annual returns with no volatility and draw down over the long run ?
C
Another scenario you may wish to consider
Trend following has actually been awful / ish for the last 10 years ? Look at the performances of funds like Dunn and JWH over the last 5 years compared to their early years ?>
Perhaps we are on the verge of a really good trend following environment where trend followers make 100-200 % annual returns ? Why not > The whole world is awash with money in short term systems that try to capture small moves and return 10 % a year
From my calculations the hedge fund industry currently manages 2.9/3.2 trillion dollars with trend followers accounting for less than 2.4 % of that - the real risk is proper volatility for the next 10 years that teaches investors a lesson they badly need to learn - There aint no such thing as 10 % annual returns with no volatility and draw down over the long run ?
C
-
- Senior Member
- Posts: 38
- Joined: Wed May 28, 2003 2:17 am
It's a good point that assignment of "good" and "bad" to particular periods requires a clearer definition of the strategy on account of the different ways there are to follow a trend. The media refers to hedge funds as returning something for a particular year but the statistic isn't very useful when there is so much variety among hedge funds.
I wonder if correlation among trend following implementations would be surprisingly low. In this thread we have at least three systems that would be considered of similar breed (mine, sluggo's, JWH's) and they don't seem to be correlated. JWH has had a sticky few years while sluggo's and mine did well. It doesn't appear sluggo's even had much of a drawdown in 1994 whereas that period is the worst case in mine.
I wonder if correlation among trend following implementations would be surprisingly low. In this thread we have at least three systems that would be considered of similar breed (mine, sluggo's, JWH's) and they don't seem to be correlated. JWH has had a sticky few years while sluggo's and mine did well. It doesn't appear sluggo's even had much of a drawdown in 1994 whereas that period is the worst case in mine.
Chris67 wrote:Cograts on being able to predict the future ?
Another scenario you may wish to consider
Trend following has actually been awful / ish for the last 10 years ? Look at the performances of funds like Dunn and JWH over the last 5 years compared to their early years ?>
Perhaps we are on the verge of a really good trend following environment where trend followers make 100-200 % annual returns ? Why not > The whole world is awash with money in short term systems that try to capture small moves and return 10 % a year
From my calculations the hedge fund industry currently manages 2.9/3.2 trillion dollars with trend followers accounting for less than 2.4 % of that - the real risk is proper volatility for the next 10 years that teaches investors a lesson they badly need to learn - There aint no such thing as 10 % annual returns with no volatility and draw down over the long run ?
C
Depending on your timeframe, you can come up with some very attractive returns using a simple breakout trend following system for the past 10 years.
The trend followers who haven't thrived recently appear to be those with a short to medium term focus, and with a tendency to trade both the short side and long.
Very long-term, long only, trend following systems have been performing very well over the past ten years, and equally well prior to this.
- Attachments
-
- 10 Year LTTF Returns.png (64.07 KiB) Viewed 25704 times
RenTech (and some others) might disagree with you about that.DeanoT wrote:The trend followers who haven't thrived recently appear to be those with a short to medium term focus, and with a tendency to trade both the short side and long.
Here is an example of one of the systems I trade which qualifies I think as short term and which goes both long and short. (see below).
But then again, someone trading your system will be able to go for long walks on the beech, enjoy the beautiful Melbourne Royal Botanical Gardens in the middle of the workweek and while away hours at Collins Street coffee shops. Whereas someone trading my system won't be doing those kinds of things during trading hours, they'll be working their ass off. Those can be vexing trade-offs.
P.S. Notice the high margin-to-equity ratio BTW. This test run is shackled by current day margins because it uses a margin filter without the benefit of lower historic margins. Still, I think this test run is sufficient to make my point.
- Attachments
-
- Example of a s/t system that trades both long and short.
- Example of a Short Term LS System.PNG (51.77 KiB) Viewed 25662 times
A little off topic there Levijean - Eric suggested that VLTTFS are the way to go but subsequent posts suggest that ain't necessarily so.levijean wrote:so, with all these backtests with jaw-dropping returns year after year, how many of us here have $1 billion already?
I think I would point to RenTech as a "best in class" example (the class of all hedge funds) and they are rumored to trade very high frequency models (much more than me I understand) and they seem to do just fine with that approach.
http://en.wikipedia.org/wiki/Renaissance_Technologies
-
- Senior Member
- Posts: 38
- Joined: Wed May 28, 2003 2:17 am
The question I aimed to raise by pointing out unlikely DD's: What are the things that improve sustainability most with trend following? The three that stand out to me are conservative leverage, diversification among markets, and diversification among strategies (which, with little evidence, I feel are not as correlated as simple logic implies).
Are there others?
Are there others?
Good old Jack Schwager found some crusty old CTA's who said the very same same things, nearly twenty years ago (link). From pp. 181-2:
and the interview goes on for seven more pages. It even mentions Nelson Bunker Hunt on p.185.So the very first rule we live by is: risk a small percent of total equity on any trade.
Second, we always follow the trends and we never deviate from our methods.
The third thing we do to reduce risk is diversify. We diversify in two ways. First we probably trade more markets worldwide than any other money manager. Second, we don't just use a single best system. To provide balance, we use lots of different systems, ranging from ...
The fourth thing we do to manage risk is ...
-
- Roundtable Knight
- Posts: 2071
- Joined: Fri Apr 25, 2003 3:33 pm
- Location: London
- Contact:
Perhaps change is the only certainty in life.
Some things in the markets may change permanently (increased particpation and increased noise perhaps)? There again, perhaps not given the latest fuss about speculation.
Some things may come and go in cycles - the attached chart may help to explain why shorting using an LTTF has been an uphill struggle this decade. Perhaps commodity prices will go up for ever. I doubt it.
Some things in the markets may change permanently (increased particpation and increased noise perhaps)? There again, perhaps not given the latest fuss about speculation.
Some things may come and go in cycles - the attached chart may help to explain why shorting using an LTTF has been an uphill struggle this decade. Perhaps commodity prices will go up for ever. I doubt it.
- Attachments
-
- CRB.jpg (33.37 KiB) Viewed 25394 times
-
- Roundtable Knight
- Posts: 2071
- Joined: Fri Apr 25, 2003 3:33 pm
- Location: London
- Contact:
A long only investor in the early 1930's may have wondered whether long only equity investment was a thing of the past. So might an investor in the 1960's (although he would have had an income from dividends, not represented on this price only chart). Both investors would have been proved wrong………eventually.
Market cycles come and go; all the more important therefore to look at as much data as we possibly can. Some “rulesâ€
Market cycles come and go; all the more important therefore to look at as much data as we possibly can. Some “rulesâ€
- Attachments
-
- Dow.jpg (24.63 KiB) Viewed 25387 times
-
- Senior Member
- Posts: 38
- Joined: Wed May 28, 2003 2:17 am
The long-only question is compelling. The obvious clue is that if something can go up an infinite amount but has a limited downside by nature, why participate on the short side if there are enough long trades to put the capital to work? It seems much more likely that "good" volatility can occur in a long trade than a short trade because the absolute magnitude of price changes tends to drop when a market is going down.
-
- Senior Member
- Posts: 25
- Joined: Tue Feb 27, 2007 10:08 am
LeapFrog, I have similar system that hold on average 2.1 days(I didn't see in your backtest the holding period), the problem with this thing is that you cannot know the exact result because you don't know the fills. My systems trades market on the open and exits also on the market at the open. There will be times when negative slippage will "balance" slippage but it's hard to tell how much... Another thing: daily trading volume is limited( for all contracts), thus with time you can't take trades like 50000 contract of Cocoa, or Cotton or even Crude Oil and hold it for 2 days. My backtest riched trillions, it doesn't mean it can be done in real trading..
LeapFrog wrote:A little off topic there Levijean - Eric suggested that VLTTFS are the way to go but subsequent posts suggest that ain't necessarily so.levijean wrote:so, with all these backtests with jaw-dropping returns year after year, how many of us here have $1 billion already?
I think I would point to RenTech as a "best in class" example (the class of all hedge funds) and they are rumored to trade very high frequency models (much more than me I understand) and they seem to do just fine with that approach.
http://en.wikipedia.org/wiki/Renaissance_Technologies