Disadvantages of uncertainty in modeling I am preparing a presentation, my work mainly concentrates on uncertainty and sensitivity analysis. I was wondering if I can convince my audience by the importance of studying uncertainty in modeling. 
I think first of stating or presenting a catastrophic effect of uncertainty, I googled about 2 days, but I don't find what I am searching for. Then I think to add some statistical facts that visualize the effect of uncertainty  in any domain including modeling like economic, environment, and mechanics. But unfortunately, I get nothing after hard searching. 
Can anyone give me an idea that supports the investigation of uncertainty, or any event where the encountered uncertainty was a burden? 
 A: Frank Knight in his 1921 book Risk, Uncertainty and Profit defined risk as known and quantifiable, uncertainty as neither. His definition has held up well over the years although some modern risk managers distinguish between quantifiable and unquantifiable uncertainty. As Spiegelhalter's article (posted in the comments) notes, the relationship is probabilistic. In my view, however, the "quantifiable uncertainty" advocates miss the boat in failing to articulate the bounds or confidence intervals around their estimates. Given sufficiently large bounds or bounds that are undefined, probabilistic "estimates" are meaningless. 
Good, current event examples of uncertainty are in fields as diverse as global warming, flood control and cybersecurity. Take the Dutch, for example. Something like half of Holland is below sea level and can survive solely due to an extensive system of dykes and levees holding back the North Sea. The Dutch relationship with the North Sea is so critical to them that they have recorded the magnitude of storm surges literally for hundreds of years. In the mid-50s, a storm surge greater than anything previously recorded innundated the country and thousands of people perished. To counteract future threats, Dutch mathematicians employed extreme value modeling to estimate the height needed for their levee system to protect them against a 1 in 10,000 year storm and they then spent enormous sums of money to rebuild this system. Today, with the threat of global warming, the Dutch are again rebuilding their storm surge control system: this time to a 1 in 100,000 year event. 
In the case of cybersecurity, the sources of uncertainty are multiple and include the unregulated nature of the internet, democritization of the skills required to make an intrusion or threat: in today's world, a teenager can push a button and hack into a network practically anywhere on the planet. Other factors include the fact that huge profits are being made from it, that the advantage is almost totally on the part of the attackers. New forms of attack are being invented constantly. Defenders against cyberattacks can do little to prevent them. In fact, the reality is so grim, that in most cases, breaches aren't even discovered until months have gone by and the biggest damage was done in the first 24 hours after the breach. Add to this the fact that these are such new forms of threat and attack that there is little or nothing yet in place in the way of international legal, regulatory, or institutional controls. The net result is that network intrusions and breaches are growing at an exponential rate which nobody has a clue how to slow down, much less stop. 
Hope that helps.
