In just one of many illustrative examples in his book "Antifragile - things that gain from disorder" author Nassim Taleb talks about the irony of sports-shoe makers touting that their most advanced models are those that most closely replicate the bare human foot. This is fascinating since it highlights that attempting to smooth out and cushion the shocks experienced during running actually had the effect of massively increasing muscular and skeletal injuries from running.
In barefoot running, the toes and arches of the feet act as shock absorbers and - crucially - allows them to be exposed to the action of running that actually strengthens them in this function. By contrast, the attempt to eliminate those specific stresses resulted in humans striking the floor with their heels rather than their toes and the additional padding robbed our feet of the varied "training" that prevents things like fallen arches and plantar fasciitus.
Another example favoured by Nassim Taleb is that of "regulators"; devices that were fitted to smooth out some of the slightly erratic running of steam engines. The idea was to make them more efficient - and for much of the time this is the effect that they had. However, on the odd occasion, the regulators would get locked in to destructive cycles of inappropriate corrections - and cause the engine to run disastrously/explosively out of control!
In fact, Taleb's book compiles a fat list of examples in which an attempt to rid systems of the natural variability (or in his words "sucking randomness out to the last drop") results in periods of apparent comfort and stability - followed by catastrophic "blow up" or "extinction" events. Those examples are drawn from an impressive array of (seemingly) totally unrelated domains from finance (stock market crashes), through parenting (minor scrapes and bruises are good for kids' development), to careers (a position in a large corporation seems more stable than being a cab driver with variable fares - right up until the corporation makes you redundant).
For me - his writings on these attempts to smooth out randomness which result in periods of stability followed by catastrophe chime very much with lots of efforts to manage flood risk by modifying river channels. There is a certain appeal to the thought that by smoothing and speeding the flow of water out to sea,the risk of it spilling out of its banks should be reduced. But is that a safe assumption to make in the built environment?? In practice that assumption can prove to be catastrophically misplaced. For instance Robert E. Criss and Everett L. Shock's paper "Flood enhancement through flood control" Geology, October, 2001, v. 29, p. 875-878 attributes the increased flooding magnitude and frequency to the engineered channelization that was designed to smooth out and reduce the threat from flooding.
Similarly, I am keen to highlight at any given opportunity the potential responses of river channels when they are dredged to produce dimensions over and above the capacity established through natural erosion and deposition processes:
Essentially - in common with financial markets and steam engine regulators - there are often hidden downsides to naive interventions in the movement of floodwater through a catchment. Those hidden downsides are generally large enough to wipe out (by many, many times over) any benefits that may have accrued during the period of relative (and illusory) stability...Again, this is true of financial trading just as much as it is true of counter-terrorism or - indeed - efforts to increase and smooth floodwater conveyance upstream of hydraulic bottlenecks...
Daniel Kahneman writes (for example in his book "Thinking, Fast and Slow) with consumate persuasiveness and good sense on how we, as humans, are predisposed to defaulting to belief in representations of the world that consist of a simple story. In fact, we are far more likely to believe something that we can form into a simple narrative than an opposing view backed up by hard, objective evidence. Consequently, we treat sharks as more dangerous than traffic - despite what we know about road death and shark attack statistics...
Professor Kahneman uses the phrase "cognitive ease" to describe the state in which we readily accept notions so long as our brains can construct some kind of story around them. It is this state that makes us extremely vulnerable to errors in reasoning (or cognitive traps) - as illustrated by the fact that we will always feel that a 90% fat free yoghurt has to be much healthier than a yoghurt advertised under the tag-line "10% of this product is pure fat". In the same way, we find it nearly impossible (unless we really force ourselves to sit down and think about it) to resist the following cognitive trap - which came to be known as "The Linda Problem":
Participants in the experiment were told about an imaginary young woman named Linda, who is single, outspoken and very bright, and who, as a student, was deeply concerned with issues of discrimination and social justice. The participants were then asked which was more probable: (1) Linda is a bank teller. Or (2) Linda is a bank teller and is active in the feminist movement.”
The overwhelming response was that (2) was more probable; in other words, that given the background information furnished, “feminist bank teller” was more likely than “bank teller. However - this cannot possibly be true; every single feminist bank teller is (of course) a bank teller - but there are lots of additional bank tellers that will not be active feminists. It is the same as saying "Linda is a red haired bank teller" is more likely than the statement "Linda is a bank teller who could have any hair colour". In other words, we all know that a 7-horse accumulator bet is far, far more unlikely than a single win on an individual race; but we don't translate that statistical knowledge well to other cases where there is an easier narrative story as an alternative.
Turning our attention to the prominent (and dreadful) recent impacts of flooding in the Somerset Levels. These events took a serious toll on the lives and livelihoods of residents - with particularly arduous circumstances being endured by those members of the farming community based on land that was inundated for the longest periods of time. However, surely we should not fall into a known cognitive trap that makes an easy soundbite - but actually completely ignores the true situation? Aren't we at risk of selling those communities at risk of flooding short by offering purely cosmetic activities?
The sound-bite of "we must dredge the rivers" so that the flood water can be taken away intuitively makes sense - and it has been readily accepted (and campaigned for). But could this be just another version of the Linda problem? Might we be accepting the simple story that we can construct in our own minds instead of taking the tough look at the complex and unpalatable evidence at hand? Consider the following, provable, facts:
In other words, it doesn't matter how many lanes you add to a motorway if there is still a road block at the far end...
All of which brings us right back round to Nassim Taleb and his writings on the concept of fragility. You see, Mr. Taleb - far from being an economic boffin with no connection to the real world - actually puts his money where his mouth is. He has amassed a (considerable) income by adopting a particular philosophy as a financial trader. His approach could be summed up in part by the following thought experiment:
It does not matter that you cannot predict which specific lorry (on which specific day) will destroy a weak bridge; it is just important to know that the bridge is fragile.
By identifying those commodities and institutions that are fragile to catastrophic blow-ups, it can be possible to insulate yourself from the negative consequences of such an (inevitable - but unpredictable in time) event. Moreover, it is actually possible to bet against such fragile institutions! This is how Taleb not only survived, but actively profited from the stock market collapse of 1987. He also correctly predicted that the credit crunch (and ongoing associated recession) was an inevitable consequence of the activities of big banking...
So, here is what we know about the significant (but obscure) downsides of trying naively and too aggressively to remove randomness to produce predictable stability within a chaotic system (such as flood water conveyance by channel modification): periods of apparent stability are followed by a catastrophic event of much greater magnitude...
Furthermore, we already realise that it is not really that important to know when a catastrophe will strike. Instead it is important to know whether we are fragile to the effects of that catastrophe when it does strike. Is the bridge strong or weak??
We KNOW that building and living in the floodplain will be, from time to time, fragile to unusually high rainfall, falling on especially impervious and steep surfaces that run into low-lying land where the tides can prevent the escape of that spectacular rainfall. It is not a consequence of insufficient intervention into the size and shape of the river channels - it is an inescapable feature of flood plains. Consequently, no measure (regardless of how expensive) can be put in place that will eliminate this risk and that fragility. However, it is interesting to note that one of the corners of Somerset that did NOT experience catastrophic flooding last winter was the area in which some of the natural variability of water escaping onto the floodplains had been re-instated...
Dredging upstream of a "roadblock" reduces flood risk?
Linda is more likely to be a feminist bank-teller?
In barefoot running, the toes and arches of the feet act as shock absorbers and - crucially - allows them to be exposed to the action of running that actually strengthens them in this function. By contrast, the attempt to eliminate those specific stresses resulted in humans striking the floor with their heels rather than their toes and the additional padding robbed our feet of the varied "training" that prevents things like fallen arches and plantar fasciitus.
Another example favoured by Nassim Taleb is that of "regulators"; devices that were fitted to smooth out some of the slightly erratic running of steam engines. The idea was to make them more efficient - and for much of the time this is the effect that they had. However, on the odd occasion, the regulators would get locked in to destructive cycles of inappropriate corrections - and cause the engine to run disastrously/explosively out of control!
In fact, Taleb's book compiles a fat list of examples in which an attempt to rid systems of the natural variability (or in his words "sucking randomness out to the last drop") results in periods of apparent comfort and stability - followed by catastrophic "blow up" or "extinction" events. Those examples are drawn from an impressive array of (seemingly) totally unrelated domains from finance (stock market crashes), through parenting (minor scrapes and bruises are good for kids' development), to careers (a position in a large corporation seems more stable than being a cab driver with variable fares - right up until the corporation makes you redundant).
For me - his writings on these attempts to smooth out randomness which result in periods of stability followed by catastrophe chime very much with lots of efforts to manage flood risk by modifying river channels. There is a certain appeal to the thought that by smoothing and speeding the flow of water out to sea,the risk of it spilling out of its banks should be reduced. But is that a safe assumption to make in the built environment?? In practice that assumption can prove to be catastrophically misplaced. For instance Robert E. Criss and Everett L. Shock's paper "Flood enhancement through flood control" Geology, October, 2001, v. 29, p. 875-878 attributes the increased flooding magnitude and frequency to the engineered channelization that was designed to smooth out and reduce the threat from flooding.
Similarly, I am keen to highlight at any given opportunity the potential responses of river channels when they are dredged to produce dimensions over and above the capacity established through natural erosion and deposition processes:
Essentially - in common with financial markets and steam engine regulators - there are often hidden downsides to naive interventions in the movement of floodwater through a catchment. Those hidden downsides are generally large enough to wipe out (by many, many times over) any benefits that may have accrued during the period of relative (and illusory) stability...Again, this is true of financial trading just as much as it is true of counter-terrorism or - indeed - efforts to increase and smooth floodwater conveyance upstream of hydraulic bottlenecks...
Daniel Kahneman writes (for example in his book "Thinking, Fast and Slow) with consumate persuasiveness and good sense on how we, as humans, are predisposed to defaulting to belief in representations of the world that consist of a simple story. In fact, we are far more likely to believe something that we can form into a simple narrative than an opposing view backed up by hard, objective evidence. Consequently, we treat sharks as more dangerous than traffic - despite what we know about road death and shark attack statistics...
Professor Kahneman uses the phrase "cognitive ease" to describe the state in which we readily accept notions so long as our brains can construct some kind of story around them. It is this state that makes us extremely vulnerable to errors in reasoning (or cognitive traps) - as illustrated by the fact that we will always feel that a 90% fat free yoghurt has to be much healthier than a yoghurt advertised under the tag-line "10% of this product is pure fat". In the same way, we find it nearly impossible (unless we really force ourselves to sit down and think about it) to resist the following cognitive trap - which came to be known as "The Linda Problem":
Participants in the experiment were told about an imaginary young woman named Linda, who is single, outspoken and very bright, and who, as a student, was deeply concerned with issues of discrimination and social justice. The participants were then asked which was more probable: (1) Linda is a bank teller. Or (2) Linda is a bank teller and is active in the feminist movement.”
The overwhelming response was that (2) was more probable; in other words, that given the background information furnished, “feminist bank teller” was more likely than “bank teller. However - this cannot possibly be true; every single feminist bank teller is (of course) a bank teller - but there are lots of additional bank tellers that will not be active feminists. It is the same as saying "Linda is a red haired bank teller" is more likely than the statement "Linda is a bank teller who could have any hair colour". In other words, we all know that a 7-horse accumulator bet is far, far more unlikely than a single win on an individual race; but we don't translate that statistical knowledge well to other cases where there is an easier narrative story as an alternative.
Turning our attention to the prominent (and dreadful) recent impacts of flooding in the Somerset Levels. These events took a serious toll on the lives and livelihoods of residents - with particularly arduous circumstances being endured by those members of the farming community based on land that was inundated for the longest periods of time. However, surely we should not fall into a known cognitive trap that makes an easy soundbite - but actually completely ignores the true situation? Aren't we at risk of selling those communities at risk of flooding short by offering purely cosmetic activities?
The sound-bite of "we must dredge the rivers" so that the flood water can be taken away intuitively makes sense - and it has been readily accepted (and campaigned for). But could this be just another version of the Linda problem? Might we be accepting the simple story that we can construct in our own minds instead of taking the tough look at the complex and unpalatable evidence at hand? Consider the following, provable, facts:
- It would be impossible to dredge the rivers to a capacity that would accommodate the volume of water that flooded such a vast area of land
- The limited "hydraulic gradient" imposed by the low-lying land (especially at high tide) significantly limits the rate at which water will run out to sea
In other words, it doesn't matter how many lanes you add to a motorway if there is still a road block at the far end...
All of which brings us right back round to Nassim Taleb and his writings on the concept of fragility. You see, Mr. Taleb - far from being an economic boffin with no connection to the real world - actually puts his money where his mouth is. He has amassed a (considerable) income by adopting a particular philosophy as a financial trader. His approach could be summed up in part by the following thought experiment:
It does not matter that you cannot predict which specific lorry (on which specific day) will destroy a weak bridge; it is just important to know that the bridge is fragile.
By identifying those commodities and institutions that are fragile to catastrophic blow-ups, it can be possible to insulate yourself from the negative consequences of such an (inevitable - but unpredictable in time) event. Moreover, it is actually possible to bet against such fragile institutions! This is how Taleb not only survived, but actively profited from the stock market collapse of 1987. He also correctly predicted that the credit crunch (and ongoing associated recession) was an inevitable consequence of the activities of big banking...
So, here is what we know about the significant (but obscure) downsides of trying naively and too aggressively to remove randomness to produce predictable stability within a chaotic system (such as flood water conveyance by channel modification): periods of apparent stability are followed by a catastrophic event of much greater magnitude...
Furthermore, we already realise that it is not really that important to know when a catastrophe will strike. Instead it is important to know whether we are fragile to the effects of that catastrophe when it does strike. Is the bridge strong or weak??
We KNOW that building and living in the floodplain will be, from time to time, fragile to unusually high rainfall, falling on especially impervious and steep surfaces that run into low-lying land where the tides can prevent the escape of that spectacular rainfall. It is not a consequence of insufficient intervention into the size and shape of the river channels - it is an inescapable feature of flood plains. Consequently, no measure (regardless of how expensive) can be put in place that will eliminate this risk and that fragility. However, it is interesting to note that one of the corners of Somerset that did NOT experience catastrophic flooding last winter was the area in which some of the natural variability of water escaping onto the floodplains had been re-instated...
Dredging upstream of a "roadblock" reduces flood risk?
Linda is more likely to be a feminist bank-teller?
Comments