Complexity’s hard Truth – the Unravelling of the World

An article about Ontonix and its technology featured in WATTNOW magazine, February 2007.

Below the full text.

Complexity’s hard Truth – the unravelling of the world

Say you were to walk into a meeting of a company or project important to you and the ambient discussion was about how things are going to be optimised for sustainable development. Faster, better, cheaper using the top-of-the-line processors to ensure the most detailed models. Would you drift off reassured by these usual platitudes, or would you find yourself in a cold sweat? If it is the latter, you probably understand something about complexity.

Intuitively most of us know that our global society has become more complex than that of our parents, grandparents, and their predecessors. We are much more interconnected, we have more options, more freedom. Our society is fuzzier yet more bureaucratic, more unpredictable, chaotic, and is likely to become even more so. Our world is definitely more complex than it was, and many of us have instinctively realised this has meaningful implications. But intuition is not science since if you can’t measure something you can’t claim to be doing serious science, and how can you manage what you can’t measure? Or, as Italian based Dr Jacek Marczyk, the aeronautics, aerospace and civil engineer who founded Ontonix, a private company that specifically deals with complexity, says, “Why do you think managing complex systems is more difficult than managing simpler ones?”

One can accurately measure a country’s monthly trade balance, the average rainfall at a particular location, how many votes a candidate got in an election, the height and weight of a person, how many cars travelled down a certain piece of road on a given day. But try making complete and precise statements summing up the economy, the climate, politics, a human being and highway traffic patterns. “To understand complexity you have to accept the conundrum – the more complex the system, the less precisely it can be measured,” Marczyk says. This is known as the Principle of Incompatibility: high complexity is incompatible with high precision.

So, how can we ever hope to quantify something as complex as our global society? Marczyk, whose company has developed a unique system, OntoSpace, for measuring and managing complexity, has done just that. He used OntoSpace to analyse raw data collected by the CIA’s World Fact Book. And the outcome is the predicted collapse of our global society sometime between 2040 and 2045. Marczyk says like with all predictions, the timing is uncertain and could be off by several decades, but the one thing he is 100% certain of is that it will happen. This projection is based on no additional modelling and assumptions, but on using the raw data and calculating global society’s upper complexity boundary. Every system possesses its own upper limit of complexity beyond which it cannot naturally evolve. Systems close to this limit are known as critically complex.

“We have seen that the yearly growth of complexity is about 5% to 6%,” Marczyk says. In essence Marczyk’s OntoSpace software determines the fragility of a system, which he has formulated into an equation: Complexity x Uncertainty = Fragility.

Or as Thomas Frey, executive director and founder of the Colorado based, DaVinci Institute, a non-profit futurist think tank, says, “As complexity of a system increases, the costs associated with it increase exponentially to the point where the costs approach infinity, and collapse is a certainty. It can be argued that every major civilisation in history such as the Egyptian, Greek, or Roman empires, as well as smaller civilisations like the Mayan Indians and Mesopotamia, each reached a point where an ever increasing bureaucracy with an ever increasing number of rules simply overloaded the administrators’ ability to comply with them, and the systems collapsed.”

Frey points out that while modern technology has given us the ability to manage systems that are far more complex than those of these ancient civilisations, he notes that following a similar curve to Moore’s Law (which predicts the doubling of computing processing power about every two years), our ability to automate has kept up with our ability to complicate. He says that the breaking point will not be the automated systems but rather the human interface where systems such as income tax in many parts of the world have gone beyond the pale of understandability and “exists as nothing more than a confusing blur to the tax paying public.”

Frey says the complexity has reached a point of being irreversible, causing the system to unravel around the edges.

The more complex a system the more functions it can perform, so we can do more than the Romans did. But our global civilisation does have a critical complexity threshold. With complexity both robustness and fragility increase. This duality – robust yet fragile – is the salient characteristic of highly complex systems. However, in the proximity of critical complexity, the system in question becomes fragile, difficult to manage, and therefore vulnerable.

Marczyk does not suggest that our global civilisation will suddenly implode, or that the complexity ceiling when reached will result in some cataclysmic wrenching apart. It could just be that a saturation point is reached. “The key message here,” Marczyk says, “is that we reach limits to all things in life. The concept of sustainable growth is a slap in the face of the second law of thermodynamics, which states that a system will always tend towards its highest state of entropy, thus highest state of chaos. Sustainable growth is not possible – the rules of complexity set limits as to the complexity that any system can naturally evolve and grow to.

The more a complex system is optimised for a specific situation, the more fragile, the more brittle it is should circumstances change. Or as the Roman proverb goes, Corruptio Optimi Pessima (when something is optimal, it can only get worse). The closer to its complexity boundary a system gets, the less robust, the more fragile it is. There are two ways a system that is close to its complexity boundary can increase its robustness. It can devolve until it recedes from the complexity boundary, or it can grow (by adding infrastructure) to increase the critical limit. In the case of companies near their complexity limit, they can drain entropy, reduce chaos, by shutting down certain divisions or they can expand so their critical complexity limit is higher – hence the enthusiasm for mergers and acquisitions. However, Marczyk points out that 70% of mergers fail, because this is an area of complexity where too many variables interact and the outcome is not a foregone conclusion.

And now the science of understand and managing complexity demonstrates its significance. Thanks to the work of Marczyk and his colleagues we can now take a system and calculate how far it is from its complexity threshold, in other words how fragile it is. We can now tell how much scope a system has to cope with unexpected shocks, how healthy the system is. In other words, Ontonix has established a radically new way of understanding and managing risk. We can now answer the question every executive would dearly like to know – how much, globally speaking, is a given corporation at risk – in other words, how far from critical complexity are we?

But before we can measure complexity we also need to understand clearly what we are talking about. “The so-called science of complexity has been around for a couple of decades,” Marczyk says, “and there are numerous research centres around the world that study complexity. The strange thing is that there is no established definition of complexity and no rational measure either! We need to define it, and dictionary definitions are not particularly helpful here. Renowned dictionaries make mistakes and confuse complex with complicated. To understand complexity you need to understand the difference.”

A mechanical watch is a very complicated system, with its numerous springs, shafts and gears which must all work precisely together. But because each component can only do one thing, each having only one degree of freedom, it cannot do anything spontaneous. It has a very limited capacity to surprise. A mechanical wrist watch is thus complicated, but not a system with a high level of complexity. In contrast take a family comprising three or four humans, spending an afternoon in a park. Predicting the actions of such a system is much more difficult, because it has a much greater capacity to surprise – the human family is a system with a much higher level of complexity. It is not sufficient to have a large number of connections between numerous components to speak of high complexity. To speak of high complexity we also need an element uncertainty – the capacity to surprise.

Numerous tools exist for simulating complex systems. These have been around for decades and use stochastic simulation, in other words based on probabilities, via the Monte Carlo method. This method, named after the famous casino because of its probabilistic nature, is the approach used to estimate the amount of water in a reservoir based on the random distribution of rainfall and water usage. The Monte Carlo method would determine what area a circle takes up within a square by sampling a random distribution of points and seeing what percentage of these are within the circle. In engineering, and other systems, where the distribution is not completely random Gaussian (Bell Curve) distribution patterns describing the probabilities of a phenomenon are used when undertaking Monte Carlo simulations.

What OntoSpace does, using results of Monte Carlo simulations, or directly any kind of measured data, is to first determine all the possible modes of behaviour a particular system can exhibit. Even apparently innocent systems can possess a multitude of modes, and it becomes clear how simplistic it would be to say that the behaviour of complex systems can be represented by just one global correlation map. The package then distinguishes hubs, critical variables that concentrate the fragility of the system. Now, when the system is a complex one with many variables, none dominant, its overall performance is determined by the collective behaviour of the many. And you don’t try and determine which variables run the show; instead you determine the hubs, the variables that connect to the highest number of other variables, in other words the nodes with the highest degree. The loss of a hub creates massive damage to the mode and possible loss of functionality. A multi-hub system is thus more robust, more resilient than a system dependant on a smaller number of hubs.

A society with more diversified hubs of political and economic power won the cold war against a society with fewer and more concentrated hubs. And conversely, in spite of all its technological, political and military might, the USA is failing in Iraq because it is trying to combat an enemy that has many hubs. A targeted attack on a single hub system may bring the system down, such as an ecosystem that suddenly crashes due to the loss of what biologists call a keystone species. The problem with complex systems such as our biosphere is we don’t know what these hubs are, as we fiddle with them, often in an aggressive measure.

Marczyk also uses his knowledge of complexity to warn against the way in which it is almost too easy for engineers to over-rely on modelling systems. “It is the path of least resistance to use existing processing power to build detailed models that refine below a level of natural precision.” One is not going to squeeze out more information than this natural precision allows, and taking into account that high complexity = less precision this particularly refers to long range climatic, political, traffic, and economic models.

“It is incongruous that one models something to extreme levels of detail, such as in case, for example, of motor vehicle safety parameters. The modelling is typically very detailed in the extreme, but the results are presented to the market as a rating of one to five stars, i.e. in fuzzy terms. There is no balance in this.”

And Marczyk points out this is not a new concept as it was Aristotle who said, “An educated mind is distinguished by the fact that it is content with that degree of accuracy which the nature of things permits, and by the fact that it does not seek exactness where only approximation is possible.”

If we are going to manage the coming complexity boundary our civilization faces we are going to have to become much more serious about understanding and managing complexity. This means giving up certain cherished ideas. “We cannot push performance to the limit. We cannot design for perfection. Sustainable growth is impossible. Everything reaches a peak and then, as much as we may not like it, decays – the universe, civilizations, and human beings.”

Every time a company executive says how that company has been optimised, understand that its risk profile has increased. At the moment we are optimising our world economy with first-world countries exporting high tech and importing cheap labour. Drives for more profit with lower investment means companies under such pressure are doing less real research and development. This seeking of maximum profit with minimum risk, minimum investment in the shortest possible time all equates to fragility. The same applies to engineering systems, which have nowadays evolved to impressive levels of complexity, sophistication and performance. At the same time, like the Space Shuttle with a faulty O ring, they can fail with surprising ease, and frequently due to very simple and trivial causes.

It was Heinz Pagels (1939-1988), the physicist and writer of several science books including The Computer and the Rise of the Sciences of Complexity who said “I am convinced that the nations and people who master the new science of complexity will become the economical, cultural and political superpowers of the next century.”

Ontonix has found a way to rationally measure complexity. “Our vision, says Marczyk, is to help his  make the world a less fragile place, to show that we need less pursuit of perfection and of detail and a more holistic comprehension”. “It could be 2025, or 2085, but our society will reach its complexity boundary, become very brittle and probably collapse.” The only hope is that it will be a soft decay and to ensure that is the case Marczyk is trying to promote the pursuit of healthier goals than pursuit of ultimate returns. “If we use the laws of entropy and complexity wisely we can have a lot, but we can’t have it all.” The world will always be uncertain, but we can make it less complex and thus less fragile.