Inserted Video
- Joined Jan 2009 | Status: Learning life... | 8,346 Posts
Memories caught in time but never forgotten
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
Niels: Yeah. .. am I right in saying that you found, actually, some quotes from some, I don't know what it was, was it a politician or something like that who actually used words that I think we use now-a-days in describing trend following?
Katy: Yes. This quote is from David Ricardo, who was a legendary political economist, and this is sourced from a book called The Great Metropolis, in 1838. He said, "cut short your losses, and let your profits run on."
Niels: Sounds very familiar doesn't it?
Katy: It's almost 200 years old, but it's the same concept.
Katy: .....where this original paper came from to give some contextual reference for that....... I was in a meeting with Hans Fahlin, who's the CIO of AP2, and Hans turned to me and said, "I don't understand. Tell me why does this (trend following) work during this period of time?" I stepped back... everybody knew that trend following tends to do well during a crisis period and I just was so irritated that I couldn't answer that question that I actually thought about it for some period of time and I went back and did some analysis and did some research and I said, "my goodness, this is pretty incredible, on certain days, when things are really bad for equity (and I'll get back to the other markets in a second) things are... something is happening." It's not just happening in equities. Most of it is happening outside of equities, and it's not just happening in commodities, it can happen in rates, or FX, or here, or there. I started looking at these days where there's these big moves in equities and I found that (and this is going to be a geeky point) .....the cumulative distribution of these particular days is actually before the distribution of the days outside......I will use an example to try and explain what this means.
.. imagine I'm trying to explain this to one of my MBA students. So if I take the days where equity markets go down, and I look at a simulated trend following system, and I take something which is called the cumulative distribution function. So how that works is you think of it as building. When you build a cumulative distribution function, it's as if you take the values and put them into a bag, and so you start collecting them. So if something has a big fat tail, you're going to see a lot of mass on the left side, and then it's going to grow up less slowly. So what you can do, actually... one particular statistical test you can do is you can look at if one dominates the other. What that means is that the cumulative distribution is farther to the left... the one that dominates to the right, and then you have another to the left. So if you have two distributions: one that is to the right of the other, completely. Then it's considered first order domination.
If you take stocks and bonds, this relationship doesn't hold, because they cross and why is this the case? It's because stocks have fat tails, which means that they end up collecting more of these worse scenarios first before bonds, but then they have much better performance latter, so their distributions actually cross. So it makes sort of a loop. But when I looked at trend following returns in a certain... some of the daily analysis that I looked at, based on a sort of a filtering rule that I use, I could find periods where there was first order stochastic dominance and I have basically never seen or very rarely seen that in financial data.
I said, "there's something here that's just different." So I thought, OK, these particular moments... something is happening where these strategies are adapting to the scenario of sort of a crisis scenario in a way that is not expected. Then if you go back and you think about the Efficient Markets Hypothesis, futures markets should be so competitive that you can't make money, right, because they're obviously the most liquid, sort of the most efficient. Then I thought, wait a minute, maybe it's actually the case that they're a little bit like Buffet, that they're liquid when others are not, so the fact that they are so liquid and adaptable and in futures is what gives them an advantage over the others in these scenarios. That's where I said, OK, so what are they getting at this period of time when things are sort of a mess? Well, they're getting alpha, because they're finding opportunities that are up and beyond the sort of normal risk measures.
So, I said, ah ha! Now I have a buzz word, it's "crisis alpha". The content of crisis alpha came out of that entire story. It originated from a question that someone asked me that I couldn't answer, and then I went back and did a research report on it, which I actually have never published but then I wrote a short article for the CME group to compliment this research paper, which was meant to sort of be for the entire industry, and that was the original paper which was meant to be for the entire industry, and that was the original paper which is called A Short Guide to Investing in Managed Futures: in Search of Crisis Alpha a Short Guide to Investing in Managed Futures, that was in 2011.
So now going back to your point about bonds and commodities. That's something that really bothered me as well, because I kept getting that question all the time. So in the book we talked about crisis alpha for commodity indices. We talked about bond crisis alpha. We talk about commodity crisis alpha, but over the course of writing this book I actually had moved more towards a new idea, and this is the idea of divergence.
.....So if you're a convergent risk taker, you have a particular view. Let's say that you view that equity markets are going to go up, as an example. If equity markets go up, you tend to take profit on that. When they go down, you'll tend to do the opposite. You'll say, wait a minute, I know that equity markets go up over the long run, this looks like a buy opportunity. I should actually double my bet, or at least hold my bet and not sell it. So in that sense, over time, when you're convergent, when you win it reaffirms to you what you believed. When you lose, it actually goes against your fundamental belief structure which is sort of a threat in some sense, causing you, in some sense, to often reaffirm your beliefs.
Katy: Starting in around 2004, professor Andrew Lo, from MIT, put forth the idea of the Adaptive Markets Hypothesis. This hypothesis is an alternative and a complement to both the world of behavioral finance, which is really a world of psychology, and the world of efficient markets, which is more sort of a physics view of the world, where you see F=MA. If you look on a spectrum, psychology and physics are very, very far apart. What Andrew brought forth, which is a really fascinating way to think about it is that markets are much more like evolutionary biology - somewhere inbetween, where the psychologist have some things to say, and the physics matters too. So if I want to give you a definition of the Adaptive Markets Hypothesis, it's an approach to understanding how markets evolve, how opportunities occur and how market players succeed or fail based on the principles of evolutionary biology. So the concept in that, based on Andrew's work, is to see the market as an ecology and to understand who succeeds and fails based on those principles. So competition will drive who succeeds; resources which are available will drive profits; and the evolution of our industry is a function of the players that are involved in the industry and the resources that are currently available. So you asked me to connect convergent and divergent. Well, what that means is that, depending on the environment, at some periods of time convergent makes sense. At some periods of time divergent, but if you want to be adaptable over a long horizon and survive, you need to be both following the herd and convergent, but you also need to be divergent so that you can innovate, adapt and sort of be more robust in times when markets are changing drastically.
Niels: Of course divergence and volatility is somewhat related, but it's not the same thing.
Katy: Yes, divergence and volatility are correlated. They're positively correlated maybe at 20%. The reason is that... I was just giving a talk about this recently for the CME, and what I said was that if you have low volatility, you tend to have low divergence, but if you have high non-directional volatility, so that means where things are going up and down, and up and down, but they're not really going anywhere. This is actually a nightmare for a divergent trend following strategy, so there's no divergence in that. It's actually low divergent. But if you have high directional volatility, then you have high divergence. So divergence is more if you take... it's basically the amount of discernable trend in price. So if you take this, sort of over an horizon and you divide the amount of movements, you're actually taking the volatility out. So if you have lots of volatility, the divergence is really the signal to noise ratio in prices. So when there's lots of volatility there's lots of noise, and therefore divergence is not very high.
It is during these specific times (market regimes) that a convergent risk taker dominates the forum. They are the ones that:
And the reason that the numbers of convergent risk takers swell is that during this regime, their beliefs are reinforced until they become 'Lords of the Domain' with a legion of loyal supporters.
It is during these times that a divergent risk taker (also called the trend follower) keeps his powder dry and avoids overly committing their diversified portfolio to trades when alpha is scarce and highly competitive. Frequently those trend followers that find they over-commit during stable market conditions, endure progressively larger and larger drawdowns and if not careful run the risk of such significant drawdowns that they are unable to recover. Now this is 'why' the diversified trend follower back-tests over as many different market regimes as they can muster as it is imperative to scale your trades and trade frequency in such a way to 'keep your powder dry' and avoid these excessive drawdowns....as 800 years of history have told us that market regimes frequently change, and during these disruptive times, trend following makes it's return in spades to recover from these protracted drawdown periods during stable efficient markets.....so typically just when you start second guessing yourself and start to think like a convergent risk taker.............an innocuous 'something' happens to disrupt the apple-cart and bring chaos to the convergent risk taker.
The prevalence of these market regime shifts are much more frequent than you might think but are impossible to predict. So while stable market conditions might prevail most of the time, market transitions occur 'some' of the time, but enough of the time to allow diversified trend followers to outperform the competition. This is what gives the fat tails to the distribution of market returns. This frequency of occurrence of market regime shifts is the reason why convergent risk takers rarely can post an audited verified track record of more than a few years and why so many curve fit EA's get sent to the sin bin.
The regime shift itself disturbs the ecology (it's participants and network of relationships that have previously emerged) and the false sense of security that have been lulled into it's participants due to previous market stability. We get panic and system disruption as fight or flight pandemonium set's in. The bulk of convergent risk takers start acting like a herd scurrying for the exits and across asset classes, and on the other side of those trades lies the 'divergent risk taker' who is simply following price action and relying on extended directional price behaviour during these chaotic periods of disruption before the market achieves a new equilibrium where then the convergent risk taker starts to emerge from their bunkers ....to repeat the cycle again and again and again.
This is when the competition for alpha becomes far easier as the bulk of previous competition are heading for the exit gate, and this is when diversified trend followers start pillaging the horde. The availability of alpha during these periods is extreme due to the 'sudden' dramatic reduction in competition and a time when the lion's share of performance for the trend follower emerge which starts paying for their enduring patience and associated drawdown legacy. Think of ecologies that have had a major disruption (such as an asteroid strike) where the bulk of the competition is wiped out and the 'meek' robust species who survives rapidly exploits the uncontested environment.....well us trend followers are the meek that will inherit the earth. :-)
Now this disruptive feature is simply symptomatic of the dynamic nature of markets. Once an equilibrium is reached, a new network of relationships emerge that create inertia and 'stick' the market into it's stable enduring state...for a period of time at least....until that small flap of a butterfly wing occurs at a strategic weak spot that sends the market into short term pandemonium.
Now the thing about complexity is that as systems become more complex, the stability of the system reduces. Think of any large complex system (like the global economy) and think what a small disruption, such as a power outage for example brings with it)........ so the future of trend following is indeed bright particularly given the nature of global markets and their current 'manipulated nature' by government intervention and the onset of a flurry of HFT. The more complex the system becomes, the greater the propensity for cataclysmic disruption. The thing however about HFT, just like mean reversion or indeed any negatively skewed strategy is that when the collapse comes, it will be sublime......so keep on keeping on and don't lose the faith as sooner or later we will be 'bringing in the sheaves'.
What we are discussing here is simply evolution in action of a complex system....... and to be a survivor and have a track record in this market of more than a few years beyond a single market transition, you need to become a divergent risk taker. This however does not preclude the trend follower from dabbling with convergent systems provided it is a supplement as opposed to the main game of his/her strategy that is uncorrelated and assists in reducing drawdown impacts. This is the preferred way to deal with uncertainty and the reason for 'how' life has evolved on this planet. Not by design, but simply by being sufficiently robust to survive in a dynamic environment. To all those diversified trend followers out there......take a bow from Charles Darwin.
C
Disliked{quote} Ive just had to evacuate anyone who was in my office due to unexceptable noise levels... What a great tune, I love this sort of vibe..I need bigger speakers ...lol.. And that's coming from someone who is in their late forties ...who said your only young onceIgnored
DislikedWeekend Ruminations - What is Trend Following and why is it a Divergent Risk Taking Approach? Below is a summarised transcript from Kathryn Kaminski (Chesapeake Capital) speaking with Neils Kaastrup-Larsen from Top Traders Unplugged Episode 041Ignored