Edward Lorenz went for a coffee break while his computer crunched some numbers on his weather forecast. Upon his return, one of the great revolutions in 20th century science began to unfold.

Lorenz, an MIT meteorologist, was a pioneer in building models of the atmosphere and running them through early computers. That day he was running his model on one of the world’s first low cost “desk” computers – the LGP-30.

The LGP-30 could be had for a mere \$47,000 in 1960 – the equivalent of about \$400,000 today. Note that I said “desk” computer. I didn’t say “desktop” computer. It weighed about 740 pounds. Bill Gates and Steve Jobs were both 5 years old.

Lorenz was using the LGP-30 to prove to the world that weather patterns could be just as predictable as the movement of planets, the occurrence of eclipses and the rising of tides. He wanted to do for meteorology what Newton had done for classical mechanics.

On that fateful day in 1960, Lorenz wanted to revisit one of his earlier simulations but he wanted to save time by not starting over from scratch. The LGP-30 was formidable but it wasn’t known for its speed.

To save time, Lorenz took the outputs from halfway through the prior run and used them as inputs for the new run. He was hoping to pick back up where he had left off.

When Lorenz returned from his coffee break, the results he found waiting for him were wildly different from the original results that he had been trying to reproduce. The two sets of results were so different that they seemed completely unrelated. They might as well have been randomly generated.

Lorenz searched in vain for a bug in his program or an error in his inputs but to no avail. Finally, he realized that the huge differences in output that he was seeing were due to very small number rounding issues in his inputs.

Rather than type in the full six digit input, Lorenz had taken a shortcut and just typed in the rounded three digit version. He had typed 0.506 rather than the full 0.506127. Lorenz had no idea that such a small change in the “initial conditions” could lead to such dramatic differences in the final output.

No one had ever realized it … and thus was born one of the great ideas of the 20th century – deterministic chaos. It also led to one of the great misunderstandings of popular science.

“Chaos” is typically defined as complete disorder and confusion. In Greek mythology Chaos is the primeval void … a place of no order.

Chaos is often associated with randomness. Both hint at a lack of pattern and predictability.

Here’s the rub, however.

Lorenz could very easily produce the exact same set of results over and over and over again. All that he had to do was use the rounded inputs (0.506 instead of 0.506127) and he could produce the exact same results that he had found upon returning from his coffee break.

The results he got were completely predictable, given the right set of initial conditions. The results were fully determined by the initial conditions. The system was deterministic. It was the initial conditions that were the source of the wildly different outputs.

Chaotic systems are predictable… if you know the exact initial conditions. If there are even small changes in the initial conditions, the final results can be so different as to seem completely unrelated.

Nowadays we say that the system is “sensitive to initial conditions.” It doesn’t mean that the system is unpredictable. It just means that it’s very hard to get a consistent output if you don’t put in the exact right inputs.

What exactly does this have to do with investing? Everything. Lorenz’s work has profound implications for us as investors.

Lorenz’s model had only 12 variables or inputs. It used only 6 decimal places of precision. It ran on an ancient computer that had less processing power and memory than the average coffee-maker has today.

In spite of that relative simplicity, just rounding the inputs produced radically different results.

Now think about today’s financial markets.

They are vastly more complex than Lorenz’s simple model.

Even if you could build a model of the economy or even a single business, how many different inputs or variables would your model require? If you could settle on a finite set of inputs, could you measure current conditions accurately enough that your final outputs would be in the ballpark of what actually happened?

Not a chance.

Lorenz coined the term the “butterfly effect” to capture this reality when he asked, “Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?”

What it means to you and me as investors is simple – prediction is hard. Anytime we find ourselves relying on prediction to increase our wealth, we had better think twice. The likelihood of our predictions being accurate in the complex world of economies and businesses is slim to none.

As an investor you should spend your time and energy managing risk for multiple outcomes. It’s liberating to embrace the chaos and reduce your reliance on prediction.

And – if nothing else – you can impress friends and family at your next cocktail party by confidently explaining that “chaotic systems aren’t inherently unpredictable … they’re just hyper-sensitive to initial conditions.”

Wishing you a profoundly prosperous 2016,