Uncertainty vs Risk When it Comes to Choices

Last week, over coffee at 7 West, Ian and I had an energizing discussion with Christie Christelis and Gera Nevolovich. The subject was our favorite topic decision-making under uncertainty.

Here is an example of a choice paradox described by Christie and attributed to Amos Twersky:

If you had a choice of (a) walking away with $100 in your pocket (guaranteed) or (b) you had a 10% chance of winning $1,000 and a 90% of getting nothing, which would you choose. Most people would go away with the guaranteed $100, i.e. risk averse behavior.

However, if you were faced with the choice of (a) definitely having to lose $100 or (b) there would be a 10% chance of losing $1,000 and a 90% chance of losing nothing, most people would go for (b), i.e. risk seeking behavior.

The question is, “Are people risk averse when it comes to gains, and risk seeking when it comes to losses?”

One way of thinking of the choice paradox is in the context of uncertainty/certainty (vs. risk), i.e., “How certain are we of a particular consequence vs. how risky is it?” The justification for this lies in the self-organizing principles from chaos theory.

We are certainty-seeking creatures in an uncertain world, and fundamental forces compel us to form bounding structures that increase order by reducing disorder; increase certainty by decreasing uncertainty, e.g. social constructs, laws, belief systems, fences, traffic signals, and on and on. Even the dotted lines on the 401.

Our cognitive decision-making processes are among our most significant natural uncertainty-reducing structures. We are neurologically wired to make decisions, and, since by definition a decision reduces alternative choices to one, uncertainty is eliminated. (Whether the decision is right or wrong, good or bad, is a different question.)

As the choice paradox is framed, the degree of certainty of the three consequences can be stated as a %. But even though expressed as a %, the degree of certainty of each consequence is a relative measure and not a probability in the statistical sense. (People actually use “probability” as a hedge in common sense reasoning)

In the first case, the choices are
100% certain to gain $100
90% certain to gain nothing
10% certain to gain $1,000

In the second case the choices are
100% certain to lose $100
90% certain to lose nothing
10% certain to lose $1,000

When asked to make a decision in each case, i.e. choose one consequence over the others, people tend to choose among the best consequences that have high certainty, and ignore any consequences with very little certainty. In the first case the best consequence – a gain - with the highest certainty is (a). In the second case, the best consequence – no loss - with the highest certainty is included in (b).

In general then, the tendency is to discount any consequences that are highly unlikely, and compare only the desirable consequences that have high certainty.

I suppose it can be concluded that, in both cases, the choosers demonstrated consistent certainty-seeking behavior. Or, in other words, there is no paradox.

The Power of Uncertainty

Uncertainty is the force that drives the evolution of reality as we know it – that is the universe of objects we can see, feel and measure. We call the evolution “change”. It can be gradual or precipitous, and involves a symbiotic relationship between uncertainty and certainty; instability and stability.

Reality depends, and always has, on information. So, which came first, information or reality?

The Moment of Truth
In that instant after the Big Bang there was nothing but a meaningless primordial soup of incomprehensibly small vibrating loops of energy called strings.

As is argued in a recent New Scientist article, at that moment when time began, there was only information in the form of energy. There was no matter, and maximum information, maximum entropy, and maximum uncertainty prevailed.

Without structures to define it, reality as we experience it did not exist.

Then the physical universe began structuring itself. Eric Jantsch in his 1984 book The Self-Organizing Universe describes the evolution. Fundamental physical forces acted to reduce uncertainty. Points of information, or strings, clumped into quantum particles. Reality, as we know it, began to emerge. Quantum particles formed into atoms. Atoms formed molecules. And so on, until macro inorganic and organic structures appeared – galaxies, stars, planets, mountains, rivers, trees, people. Finally, social structures formed to further reduce uncertainty – families, tribes, cultures, villages, cities, nations, shops, factories, corporations.

Meaning came with reality. Fact as truth came with meaning.

Information was lost, but uncertainty was reduced. Time became the operator. The self-organizing principles of chaos theory held.

First a slight, but relevant diversion.

A Theory of Everything
The emphasis in Janstch's book is on chaos theory and a rationale for complexity. However, implicit in the description is the need for a unifying “theory of everything”. That is, an explanation that blends relativity, quantum mechanics and general thermodynamics. Brian Greene provides this in his book on string theory, The Elegant Universe.

It is stellar-mass black holes that have thrown relativity and quantum mechanics a curve. These black holes are objects in our universe that exercise a gravitational effect so out of proportion to their size – zero volume and infinite density - that not even photons can break away. Both relativity and quantum mechanics should hold.

Relativity describes gravity and the behavior of large objects as predictable. Quantum mechanics describes the behavior of very small objects as random and unpredictable. Since one description contradicts the other, Einstein was right and a one-size-fits all theory is an imperative. String theory is the latest and most promising attempt by physicists at a unified theory.

Next Posting
Back to uncertainty, information, reality and self-organization.

More about Uncertainty

Last Friday, Burhan Turksen – from the engineering department at UofT - Alaleh Azad and I discussed uncertainty over a wonderful lunch on Baldwin Street. This morning, looking back on the blogs already posted, I see that uncertainty is indeed the reoccurring, underlying theme in decision-making. It warrants a closer look.

So, in theory and practice, what is uncertainty? Chaos theory begins to inform the answer.

We travel through space and we travel through time. The first determines where we go – to the office, to the grocery store, to the cottage, to Mars. The second determines what we become – humans, thinkers, mothers, fathers, Canadians, bankers, soccer players.

The Certain Journey
The journey through space is reversible, predictable, and – barring any obstacles - linear. The classical Newtonian laws that govern it are well understood - simple, invariant, and universal. Each point remains constant in the geometry of space, and we move around among them. Gravity drives the journey. Mathematics calculates it. Science measures it. Technology enhances it. In principle, certainty prevails.

The Uncertain Journey
It is the second journey – our evolutionary trip through time - that is complicated and uncertain. It is irreversible, beyond our ability to predict, and hard to calculate and measure. It is governed by laws that are not simple, invariant and universal. It occurs concurrently at many levels of description – quantum, molecular, and the macro levels that we can see. Different laws hold at different levels.

At the quantum level the journey is a puzzle. Quantum particles can be in different states and different places at the same time. Their behavior is deterministic but unpredictable. The laws that govern this level defy the classical laws of physics. Instead, Schrodinger’s equation and Heisenberg's uncertainty principle seem to hold.

We observe the molecular level through microscopes. Chemical, electromagnetic and genetic descriptions suffice.

At the macro level the journey is governed by thermodynamic laws. Not the Second Law which describes the devolution of closed systems, but the General Law - discovered by Nobel laureate Ilya Prigogine - which describes the evolution of open systems over time.

From Chaos to Certainty
Chaos theory is the offspring of the General Law, and the self-organizing principles derived from the theory point to the role of structure – physical and symbolic - in reducing uncertainty. We build fences, skyscrapers, belief systems, norms, legal systems, languages and mathematical formulas in order to produce a degree of certainty in a complex, confusing world.

What are these principles of self-organization? And how do they act to reduce uncertainty?

Here we go again

Decision-Making
A year ago, in a flurry of enthusiasm, I started this blog. Now, a year later, I take up the stream of consciousness again. The subject of choice is still the same – decision-making in general and extreme decision-making in particular.

Rereading Penrose’s 1989 book, the Emperor’s New Mind has inspired me. Decision-making is explicit or implicit in everything that Penrose claims, and the book is interesting both in what it anticipated and what has transpired in the world of computational intelligence since.

The Quintessential Cogntive Function
Needless to say, we make thousands of decisions a day. Our survival depends on it. This makes decision-making the quintessential cognitive function. Fortunately, nature has equipped us with a complex structure – our brain – to do this.

Our brain takes in, sorts, stores, remembers, and processes a constant flood of sensory and symbolic information from our external and internal worlds. It then draws conclusions, makes decisions, and directs our actions. Why is this important?

Decisions and Uncertainty
In an uncertain world, we seek certainty. Decisions provide it. They eliminate alternatives. Reduce choices to one. Do/don’t. Yes/no. And, for better or worse, they commit the decision-makers to a certain course of action. Either you sleep in or you don’t. Either you buy the stock or you don’t. Either you take the job or you don’t.

In the face of uncertainty, as conventional wisdom has it, “Any decision is better than no decision.” But,

“What are the limits to our decision-making powers?”

“What about conscious and unconscious decisions?” “Can we make more than one conscious decision at a time?” And if we can, “How do we do it?”

“What is the difference between good decisions and bad?” “Between right decisions and wrong?”

“Are we the only intelligent agents around?”

Limits to Decision-Making Powers
What about the limits to our decision-making powers? Decisions come in an infinite assortment of shapes and sizes. All are self-similar, but at the same time each one is different. No single decision fits all situations. Each individual decision must be made on its own merits and in its own time. This makes decision-making a never-ending process and stretches our decision-making powers in many ways. This begs the other questions, and opens up a huge area for future discussion.

Lorna and Lotfi Zadeh

More About Extreme Decision-Making

Extreme decision-making is a subset of the universe of decision-making, i.e. the personal and professional decision-making everyone does 24/7. The following comments on corporate decision-making in general. It is going to lead us to a discussion of extreme decision-making by hotshots in the corporate world.

Corporate Decision-Making
For better or worse, everyone in a corporation is a decision-maker. Ideally, the decisions they make should be the right ones at the right time - arrived at rationally with rigor and consistency.

However, decision-makers tend to “just do it”. This makes corporate decision-making inconsistent, hard to justify, and vulnerable to moods and the time of day. All a corporation can realistically hope for is that more right decisions are made than wrong – with a minimum of bad ones.

Corporations put their decision-makers under pressure to maximize the right decisions and avoid the bad ones. And, to do it in a timely fashion.

Decision-makers get some help from existing IT. There are technologies that collect, store and retrieve vast amounts of data. Some that mine databases. Some that calculate probabilities or crunch numbers. Others that make simple binary inferences. Still others that deliver images in amazing graphical displays.

However, at that critical point in the process when professional judgement is required to aggregate the evidence, draw the conclusions, and make the right decision, decision-makers are still on their own. They are left to make the calls in their head without the benefit of technology. It is the point at which the risk to the corporation is the greatest.

DecydeWare™
DecydeWare™ is new intelligent decision technology that helps decision-makers make sound, timely judgement calls. At its core is an inference engine based on a patented algorithm that models the way exemplary decision-makers reason.

DecydeWare™ can be used wherever professional judgement is required. It is easy to use, and makes corporate decision-making more transparent, consistent, and mathematically rigorous. It improves productivity, while reducing the pressures and risks inherent in decision-making.

DecydeWare™ is available on a network appliance under the trade mark DWENA™, or as a stand-alone application. It can be customized for any decision-making situation. It integrates seamlessly with legacy infrastructures. It adds value to existing decision technology.

By improving corporate decision-making, DecydeWare™ improves the bottom line.

For more information call 416-534-0738, or 416-488-4972

What is this blog about?

This blog is a stream of consciousness about extreme decision-making. That is, decisions you make when time is short; or you don't have much info; or it is an unfamiliar situation; or the stakes are high.

Decisions where instincts, intuition or gut feelings kick in.

It started with the Fortune 75th-Anniversary Special Issue on how to make great decisions. And, will fill in a lot of the missing pieces.

Like, "How do you really make great decisions?"

Or, "If there is a clear difference between a bad decision and a wrong decision, what is the difference between a good decision and a right decision?"

"Can a good decision be the wrong decision?" Or, can luck be a lady, and a bad decision becomes the right decision.

The next posting will start trying to answer some of the questions.

This blogger is a chaos theorist and fuzzy logician with a PH.D. from UC Berkeley. But, don't hold that against her.