20/11/2008 Cropped

About a year ago I suggested I might post a fractal image each Friday. What was I thinking?

Well, a combination of guilty conscience about a commitment unkept and this sentence in Didier Sornette‘s cheerily entitled book Why Stock Markets Crash: Critical Events in Complex Financial Systems prompted me to revisit this partial promise:-

It turns out that many of the natural structures of the world are approximately fractal and that our aesthetic sense resonates with fractal forms.

Those who remember my misdirected concern about dangerous trees may appreciate that the oak has been safely pruned, and the only objects falling now are the autumn leaves and occasional acorn.

My recent routine interest in trees, and flora in general, seems closely correlated with a) the acquisition (for no financial outlay) of a Nokia N95 mobile phone containing a 5 megapixel digital camera and b) adherence to the paleo diet. The latter, you might think, is not seriously possible. But putting aside the confirmation bias, it has not been the only manifestation lately of a heightened sensitivity to fractal forms. Spooky.

More, if you can bear it, at my Flickr Photostream.

Donate and help me buy back my Fender ('About' tells you why)

Tags: , ,

I had a migraine a few weeks ago. As migraines go it was a breeze, really. There was no piercing headache, just a vice-like tension that I would normally associate with the before- and, to some extent, the after-effects. It was not a muscular tension-headache nor alcohol-related.

I didn’t experience the normal visual aura, but I’m sure it was a migraine because it was preceded by a strange faintness accompanied by shifting vision. I felt funny for most of the following week, then had a severe headache last Friday. The previous day I’d worked out hard in the gym. So I took medical advice. They said it was probably a virus that had triggered the original migraine; it was not surprising for the symptoms to be there a week later. Take paracetamol, they said. You’ll be fine. And so it turned out.

365 Days - Day 208 - Migraine

I’ve had one or two really severe migraines, but I’ve been very lucky. Apart from the couple in my teenage years that would conform to the agonising archetype blighting many people’s lives, I’ve been a light sufferer, by any measure. For 10 years of adulthood I had none. Then just a handful with no pattern or recognizable trigger. And some of those were easily knocked on the head by the early ingestion of paracetamol.

Most recently, my first half-marathon triggered one. And after the London marathon in 2005, I succumbed. In both cases my training had been incomplete: I’d overstretched myself. That would constitute a huge stress: an obvious trigger. I’ve wondered too if the demand for calcium/magnesium following that excessive hammering on bones and joints might not have helped. Too much to know.

Because I’ve escaped lightly, my knowledge of the significant progress of migraine science has been almost (but not entirely) non-existent. I noticed only yesterday, for instance, that Oliver Sacks wrote a book called Migraine in 1970 (with a revised edition published in 1992).

But the subject hove back into my view for a couple of reasons recently. I learned, for instance, that migraine as a neurological condition has a close genetic alignment with epilepsy, and that migraine sufferers are more susceptible to background noise; there is a similar phenomenon (which I don’t fully understand) in relation to eyesight.

My late brother suffered from epilepsy. But, because his epilepsy was apparently brought under control by a similar progress of pharmaceutical research, the family had been largely able to forget the underlying seriousness of a condition which reportedly affects 60 million people worldwide; migraine, by contrast, may affect as many as 300 million. The general impression — certainly one that my brother held — was that the major risk to his health came from the long-term effect of such chronic drug dependency on his vital organs. But it seemed that we could feel blessed that it wasn’t a whole lot worse; going back a generation or two, some family members’ lives had been completely wrecked, and this had chronic knock-on consequences for their carers. These days there are still those whose condition does not respond to treatment.

So, when I was researching my earlier post on Didier Sornette and the housing market, I came across a presentation he delivered in Oxford in January. It revealed some of the wider applications of complexity theory beyond the geophysics where Sornette started. In collaboration with several others, Sornette has published a paper or two explaining how the study of data sets of the brain activity of epileptics (specifically those whose condition does not respond to drugs) showed patterns akin to seismic data of earthquake incidence. The hope is that this might lead to some better method of prediction for sufferers.

The maths is rather intimidating and I’ll try to paraphrase the following as I go along, and link to definitions. Fingers crossed:-

That the pdf [probability density function] of SZ [seizure] energies E follows a power law, and more importantly that its exponent is beta almost equal to 2/3 (as for EQ [earthquake]), has far-reaching, statistical-clinical implications: the mean and variance of E are mathematically infinite, which means in practice that the largest SZ in a given time series controls their values (3). As a consequence, variability is dominant and “typical” has no meaning. The energy pdf, and specifically its heavy tail, also suggests an explanation, at a mathematical-conceptual level, for the proclivity and capacity of the human brain to support status epilepticus, a potentially fatal condition characterized by prolonged/frequent SZ during which the brain does not return to its “normal” state, even when SZ activity abates.

Well, I think the key point is ‘variability is dominant and “typical” has no meaning’, which we liberal artists would tend to capture with the expression ‘to be in a constant state of flux’ although that does not quite cover the sense of unbounded potential for an extreme spike. The problem is that we like to convince ourselves that there is some “normal”, some stability. And, on the surface, so it may appear.

And then Sornette explains:-

In seismology, it has been recognized that the many small, undetected EQ provide a major if not dominant contribution to the triggering future of EQ of any size (7). Prolonged recordings of brain cortical electrical activity (ECoG), the equivalent of seismographs, from epileptic humans and animals contain frequent, low intensity, short bursts of abnormal activity unperceived by the patient and observers and interspersed with infrequent, but longer, more widespread, and more intense bursts (convulsions) (4). The SZ-EQ analogy, including the evidence presented here for an inherent capacity of SZ to trigger future SZ, suggest that a workable prediction scheme should use the triggering by, not only past perceived (clinical) SZ, but also the myriad of unperceived (subclinical) abnormal neuronal bursts.

Sornette’s applied work highlights the cross-disciplinary relationship of the science of complexity and reminds us too that some part of our population suffers from extreme non-linearities in their day-to-day lives. And how more vivid can it be, as the picture above shows, than the fractal manifestation that is the migraine aura; when I first used to experience it, it would mark the beginning of hours of debilitation.

Meanwhile, news was reported a few weeks ago that untreatable epilepsy in children responded to a high-fat/low-carbohydrate diet. This is not particularly new. I notice now that Mark’s Daily Apple picked up a Science Daily report back in January to similar effect. Mark was also the original pointer to the picture above (thanks again Mark). So-called ketogenic diets have been known to be effective in treating even the worst sufferers from epilepsy as far back as the 1920s. And I even found a 1910 medical report quoted on a low-carb forum where a doctor had noted high levels of candy consumption among two chronic epileptic sufferers, one adult, one child, he’d been asked to treat. Drug therapies became preferred later because the higher-fat diets were found to be more difficult to follow, ostensibly for cultural reasons. Today there is a shortage of dieticians to help apply what you might call a clinical diet, where each gram of carbohydrate is very closely measured.

Well, it makes more and more sense to me that diet — and our modern carb-laden diet — has much more to answer for than we allow when we think that we are eating a “healthy mixed diet”. But it’s a struggle to remove easy grain-based carbs, and one has to wonder whether it is a sustainable option for the planet as a whole. Since I can afford it, I’m making the switch, but mainly because of the evidence that grains may play a role in activating cancer genes. I can’t ignore those pointers; I’m 43 and since January the oldest surviving member of my immediate family. Both my parents lived ostensibly healthy lives. That alone should predict that at least one would still be with us since my grandmother was alive just 7 years ago at 91.

Because of the complex, fast-moving chain of events that led to my brother’s death in January, it was hard for the surgeons to provide the family with a satisfactory narrative. I missed the chance to speak in person with the clinician; I was racing down the interstate in (melo)dramatic fashion in order to arrive before what turned out to be a technical pronouncement of death. My brother’s state on arrival in hospital the previous day was not materially different from when I arrived 10 minutes after the certification; he was on life-support simply for the purposes of organ-donation.

But that does not matter. What was described was a total neurological event — a seizure that affected all his vital functions. SUDEP — or sudden unexplained death in epilepsy — is what I understand it to have been, although that was not the word the doctor used. Or perhaps it was status epilepticus.

It’s ironic that neither SUDEP nor status epilepticus was something we knew about beforehand, or ever discussed as a possibility within our family. As I said, my brother’s principle pre-occupation in terms of epilepsy-related health (based on counselling I assume he received quite early on in his life) was that his long-term health would be compromised by the medicine he took rather than the diet he was exposed to. I think it made him fatalistic. I have no recollection that a low-carbohydrate diet would have improved his teenage outcomes, and it was something my mother would surely have responded to, had she known.

Later in life, my brother ate a standard North American diet, and there is no suggestion that this was a contributory factor to his SUDEP, but I have to wonder. Not least because the science of low-carb and the science of earthquakes both point to epilepsy from different perspectives. And scientists like Sornette and De Vany are using the same maths across these various domains.

Changing diet may not be a panacea, and I may already have sown the seeds of my own demise, but you don’t not pay into a pension scheme because you didn’t pay in before. That sort of fatalism does lead to literal and metaphorical penury. But above all else, these findings all suggest that a lot more critical reporting should be applied to questions of public health, preventative medicine, exercise, diet fads and even agricultural subsidy. That obviously ain’t happening at the moment. Indeed, the recent coverage of the ketogenic diet in the BBC/Lancet does not consider whether a lower carb diet contributes to a reduced risk of seizure more generally, and therefore might act to forestall a sufferer reaching the kind of tipping point that Sornette’s science is point toward. It is dealt with in the specific of untreatable epilepsy with no extrapolation that more general metabolic risk factors need to be considered or highlighted for all sufferers.

Well, when I asked in one of the leading cookery shops with a vast, if not complete, array of cookery titles if they had anything on the paleo diet, they had no clue what I was talking about, unsurprisingly. So there is much work to be done. That said, Nassim Taleb‘s advocacy in The Sunday Times the other day certainly has led to more Googling of “paleo diet” and other associated terms, from what I can see here, including searches for Prof De Vany.

Photo credit Auntie P @ Flickr (CC)

Donate and help me buy back my Fender ('About' tells you why)

Tags: , , , , , , , , , ,

Brown Bear, Brown Bear, What Do You See?I was thinking about the geophysicist Didier Sornette the other day. The reason being that (in my counter-factual way) I wondered what the world would look like if research (or a prediction, or an analysis) by people like Sornette were avidly watched — front-page news even. And then I remembered that I’d already written that post a long time ago, in my first blog. Like the other Yogi, it was déjà vu all over again.

Every day that I wake up to more bad news about the credit crunch, I feel slightly nauseated. It’s a bit like when you’re on a boat and the weather is closing in. Or the point that night falls and you’re out of sight of land. Or both. You’ve been there before, but your night vision needs to kick in. Time to hit the chart table, fix your position, re-evaluate how much sail you are carrying. A combination of nerves and trepidation focuses the mind. The concern is not so much for yourself, but for others. You’re in a complex system. Your fear must not guide you. You need to be confident, but careful. You may be master of the ship, but not the elements nor the other seafarers. A wetted finger is not good enough to figure out which way the wind’s really blowing. You need to calculate and apply learned heuristics, the wisdom of ages, one of which is “don’t rely on electronics”.

Finance being the bad-news-of-the-day for months on end is something I’ve never experienced before. I cut my teeth as a journalist during the extended bear market in oil that ended with the First Gulf War. The build-up was virtually a private affair for those of us who were specialists; no-one cared that much that the economy was benefiting from lower oil prices. The inflection point, when it came, was very public and geopolitical. Its consequences are still being worked out. I certainly did not see the Iraqi invasion of Kuwait coming; I was convinced that Saddam Hussein was just posturing. And yet, in retrospect, I vividly remember a conversation I’d had with a wise soul from the Middle East who showed an inexplicable agitation a few weeks before the invasion that you might characterise in the same way that animals are said to become jumpy before an earthquake. What was upsetting him was that he could read the runes whereas his colleagues could not. I was 25, still working below decks, lucky that he would take my call, and lacked the experience to fully engage with what was bothering him. The build-up to the credit crunch has been different and in some ways has already engaged the entire economically active population in psychological and also very concrete ways. The fall-out looks like being just as comprehensive.

It would be helpful if one could feel some sense of vindication, but it just ain’t happening. When you see someone driving recklessly, you don’t know whether it will end in a crash; that you are on the same bit of road — to mix metaphors — means you are inescapably in the same boat.

Getting it wrong is the sine qua non of economic forecasting. As the Stand-up Economist (whose gig on Saturday night at Oxford’s OFS I’ll be attending), says:-

Micro-economists are people who are wrong about specific things, and macro-economists are wrong about things in general … macro-economists have successfully predicted 9 out of the last 5 recessions.

Weather forecasters are often pilloried for getting it wrong. But, of all specialists, behavioural studies have revealed that they are the least confident in their own predictions. Economists and stock analysts, by contrast, are the most cocksure. And yet, the same people who failed consistently to identify the scale of the danger are also asked now to explain what happened. Nice work if you can get it.

Predicting markets is a notoriously tricky business, arguably foolish, and the great criticism that bulls usually level at those bears predicting bubble-bursts is that “even a stopped clock is right twice a day”. But what’s the inverse, exactly? Can’t the same criticism be levelled at the bulls? — precisely how “right” are they the majority of the time? And what’s the consequence of the bulls being very wrong just the once? Think back to Joanna Lumley playing Purdey [sigh] in The New Avengers in the 1970s, having to shoot her way through a kind of paintball training course. She was pleased that she’d scored 99%, marked by a single red dot that represented a bullet. Her sidekick, Gambit, pointed out it’s the 1% that kills you.

But I take the stopped clock thing seriously as a criticism of scepticism because it disparagingly suggests inaction and risk-aversion — who would want to be Chicken-Licken, after all? Certainly not the Emperor With No Clothes — and it cropped up in a piece of newspaper coverage about the credit crunch recently. My James Cramer reference the other day bears some reflection too. He would maintain, I believe, that his spiel is aimed at those with spare cash to gamble. But I think, in truth, he has been a cheer-leader for an industry that has been sailing toward the storm carrying every last scrap of sail in the locker.

But then, there is a problem with consistency. It’s generally over-rated. The ability to change one’s mind without shame should be more highly prized. As should be the ability to accept, without regret, that things may turn out better than one fears. Many a fisherman decides to stay in port only to find his catch and income is lost to a storm that doesn’t quite descend. He takes risks for a living, but I’m sure has learned too that it’s better to be wise before the event when so much is at stake. By contrast, a lifeboat man will put to sea in all weathers. But you’re taught at navigation school that he’s not to be confused with the AA man who will come and fill up your tank if you run out of petrol; he should only need to put to sea for the real black swans, not your incompetence. He won’t make that judgement, of course, but will respond to your Mayday anyway.

And so I was looking back at my own adventures as a Jeremiah, thinking about questions of timing. And that’s when I remembered that old post from my earlier blog about Didier Sornette. Sornette has long been on my reading list and in my view is one of the larger anti-heroes of modern finance that comprise my anti-library of unread books. He fits into that category where the Econophysics blog sits. The jacket of his book on markets, like Mandelbrot’s, shows fractal snail-shell patterns: you get the picture. I must buy it some day.

But I did read one or two of Sornette’s papers when they came out. I found them compelling, although the maths was completely impenetrable for me. It would be hard to find a more serious analysis of how vulnerable the markets had become at that point. That was the time to take in sail, batten down the hatches, and prepare (if necessary) to trail warps, spill oil on troubled waters, consider the possibility of removing all sail — what ocean yachtsmen call “bare poles” sailing, in the case of the perfect storm. In finance, it would mean de-leveraging early, not now.

At a very small talk I attended with Nassim Taleb in London way back in 2004, Nassim was asked by a London quant whether he thought the UK property market was in a bubble. Typical of Nassim at that time, I believe, he was confessing to not reading the newspapers so had no idea. The quant persisted that Sornette thought UK housing was in a bubble. Taleb’s response, if I recall correctly was this: “If Sornette thinks there is a bubble, then there is a bubble”. These are things I tend to remember.

Interesting, because Alan Greenspan was defending himself in the financial press the other day — and has done many times before — saying that it’s not possible to identify when markets are in bubbles. It’s a view that the prediction industry likes to repeat. But my understanding of Sornette’s science is that this is just not correct; you can identify bubble conditions from within trading price data using the same approach a seismologist does to gauge the susceptibility of the fault lines between tectonic plates to a sudden shift. I think the mathematical model he applied to the housing markets goes by the name of “log-periodic oscillations”. Predicting when the quake will occur, and with what magnitude, is the problem. That is still a work in progress, but one guesses that Sornette will be at the forefront of it as it unfolds.

Anyway, this is part of what I posted way back in June 2005 in my first, rather arch attempt at blogging called “Not that I’m Biased”:

If one is looking for a truly disinterested expert, and one with the latest knowledge on bubbles, we recommend geophysicist Didier Sornette. The mathematics of Sornette’s discipline is well beyond the lay reader. The essence of it is to show how complex systems work. He is an expert in the study of earthquakes. Stock market and housing crashes are the financial equivalents.

When people think about housing they don’t tend to think of a complex system. They will first think about their own house, those in the neighbourhood, and then a national price index recently described in the press which provides a sense of overall direction. They will probably then invoke a sense of someone who made a killing on property, or whom they saw renovate and sell at a profit on some TV show. From this they will make decisions to buy or sell. There is a strong element of imitation in what motivates them.

These behaviours are definitely part of what makes up a market, but Sornette’s specialism is in analysing them mathematically through study of the price activity of markets. Sornette’s last paper on housing demonstrated that the UK housing market would peak late 2003 or mid 2004, and then be susceptible to a crash. At that time, he did not characterise the US market as a bubble, but in his latest paper he shows that, two years on, the US is in a bubble.

A bubble with a crash in the UK will be one thing, but a serious reversal in the US would be very damaging. It remains to be hoped that the pump priming that occurred in 2000-2001 has not created a greater problem from which the world economy will suffer a more severe hangover.

It is no doubt a symptom of our collective aversion and lack of understanding of mathematics that Sornette’s work is not major news.

When the Asian tsunami hit, there was much hand-wringing about why the cooperation required to create an early warning system had failed. And yet, we know that events of that size are indeed extremely rare black swans, even though they appear to live in the folk memory of some of the coast-dwellers on Africa’s eastern seaboard. Financial shocks are coming with increasing frequency, but financial institutions, governments and the regulatory authorities — let alone the critical faculties of the media, — do not seem to be prioritising really listening to the complexity folks, despite the increasing volumes of accessible literature they have been generating in the past several years. It’s something I discussed yesterday with Brooke Harrington from the Max Planck Institute after her talk at the RSA. But that may have to wait for another post.

Photo credit: BrittneyBush

Donate and help me buy back my Fender ('About' tells you why)

Tags: , , , , , , ,

In the past few weeks the US press has been full of talk that the real estate market is experiencing a bubble. Even Fed chairman Alan Greenspan (who no doubt caused it by keeping US interest rates so low for so long) has pronounced on it. If the UK experience is anything to go by, it could be said that the day before Greenspan spoke was the peak. House prices in the UK have been in effective retreat ever since Bank of England governor Mervyn King warned in the late spring last year that anyone buying a house should be prepared for prices to fall. The next day buyers went on holiday.

If one is looking for a truly disinterested expert, and one with the latest knowledge on bubbles, we recommend geophysicist Didier Sornette. The mathematics of Sornette’s discipline is well beyond the lay reader. The essence of it is to show how complex systems work. He is an expert in the study of earthquakes. Stock market and housing crashes are the financial equivalents.

When people think about housing they don’t tend to think of a complex system. They will first think about their own house, those in the neighbourhood, and then a national price index recently described in the press which provides a sense of overall direction. They will probably then invoke a sense of someone who made a killing on property, or whom they saw renovate and sell at a profit on some TV show. From this they will make decisions to buy or sell. There is a strong element of imitation in what motivates them.

These behaviours are definitely part of what makes up a market, but Sornette’s specialism is in analysing them mathematically through study of the price activity of markets. Sornette’s last paper on housing demonstrated that the UK housing market would peak late 2003 or mid 2004, and then be susceptible to a crash. At that time, he did not characterise the US market as a bubble, but in his latest paper he shows that, two years on, the US is in a bubble.

A bubble with a crash in the UK will be one thing, but a serious reversal in the US would be very damaging. It remains to be hoped that the pump priming that occurred in 2000-2001 has not created a greater problem from which the world economy will suffer a more severe hangover.

It is no doubt a symptom of our collective aversion and lack of understanding of mathematics that Sornette’s work is not major news.

Donate and help me buy back my Fender ('About' tells you why)

Tags: , , , , , , , , , , ,

the knackered hack

Tim Penn
Alltop, confirmation that I kick ass

free updates by email

t-shirts for tired writers

Support This Site

knackered eye view

www.flickr.com
This is a Flickr badge showing public photos from knackeredhack. Make your own badge here.

Kino’s Viktor Tsoi

Kino's Tsoi
Close
E-mail It
Socialized through Gregarious 42
Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported
This work by Tim Penn is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported.
make PrestaShop themes