Raisi Is Dead - A High-Risk Low-Probability Event
In 2009 I began to use social media to document events in Iran during the "Green Movement," the popular unrest following the rigged election of Mahmoud Ahmadinejad. At the time the methods I was using were revolutionary and virtually unused by mainstream media organizations. A small collection of bloggers and hobbyists realized that we could track events that were unfolding in a nearly-complete media blackout zone using Twitter, Facebook, and Youtube. We would collect, verify, and analyze thousands of social media accounts to put together a bigger picture, one that was often missed by field reporters who could only see one street at a time. The field reporters also played a vital role, however, by giving us a reliable narrative as well as crucial insight that is missed when just using social media alone.
When the Arab Spring uprisings of 2011 happened, I took these techniques and honed them to the point where the small team I was working with had a better understanding of conflicts like Syria or Bahrain than many Western intelligence agencies, and many people in several of those agencies told me that. By the time my team at The Interpreterwas covering events in Ukraine and Russia in 2012-2017, we were regularly cited by the biggest news agencies in the world.
The basic idea is that all those social media posts, if verified, can be data sets. No longer are we reliant on a single person -- either a source or a journalist -- for information. We can synthesis a huge body of data, even giving numerical and statistical insights where before we were reliant on human ones alone. We could measure, for instance, how many protests there were, in how many locations across which geographical regions, with how many people. We could measure the sentiment of those crowds, or sometimes even the racial and religious makeup of them. We could count tanks, identify weapons, assess battlefield victories and losses.
But during this time I became intrigued by assessing probability of events happening. Could we measure a protest, and the crackdown against that protest, and use it to estimate the probability of a protest movement turning into a revolution? Could we use our methods to monitor a revolution to determine the probability of a regime falling to the rebels? During this work I also began to understand two things: low-probability eventshappen all the time, and high-probability events fail to happen.
Sometimes these irregularities happen because we've failed to properly assess the likelihood of how certain actors will behave. For instance, it was obvious that Syrian President Bashar al-Assad would do anything to stay in power. But when so many of his forces were deserting at such a high rate, it was less obvious that others would stay. It was obvious that Iran and Hezbollah would step in, but it was less obvious that the world would allow Assad's airforce to genocide the Sunni populace without at least setting up a no-fly zone.
Other times we're working with incomplete data. In 2014, civilian airliner MH17 was shot down by Russian-backed forces who had invaded eastern Ukraine. Many were shocked that high-powered anti-aircraft systems capable of shooting down such an aircraft were operating in the area. I had written extensively that they were but was shocked that civilian airliners were still flying over these battlefields. These two data sets needed to be combined to properly assess the risks.
But sometimes the craziest things just happen. This morning we are greeted with the news that Iran's President Ebrahim Raisi was killed when his helicopter crashed. Raisi was not only president but was a leading candidate to replace the Supreme Leader, Ali Khamenei, who is 85 years old and in bad health. I guarantee that nobody had "Iran's president and candidate for Supreme Leader dies in helicopter crash" on their bingo cards yesterday, and yet here we are -- a low-probability high-risk event has happened. And each time an unexpected event happens it triggers a series of dice rolls, each one with its own probability and risk. Is there a popular uprising? Does the Iranian government crack down on dissenters even further than it normally does? Does this make the probability of popular unrest even more likely once the Supreme Leader dies since a leading candidate is out of the running? Does Iran blame this incident on a foreign power or terrorist group?
As a nation we are trending towards isolationism. Many on the right and left think we should pull back from world affairs. This is a mistake. The world is full of risk. We've been able to manage a lot of it so Americans don't even see it. When we pull back from the world, the low tide exposes all the sharp edges the water hid -- and it stinks. Furthermore, when low-probability high-risk events happen, we won't be prepared to react.
