Mr. Markov's Merry-Go-Round

We live in an age of invisible giants. Complex systems are present in all aspects of modern life, and, although we typically flow through them without issue, many of these systems can crush human lives when they fail.

Commercial aviation is an obvious example, but so are food and drug safety, automobile safety, our energy infrastructure, the water supply, and countless other processes at work behind the scenes.

For each of these critical systems, various public and private authorities have enacted regulations, best practices, failsafes, and standards that are meant to prevent life-threatening failure modes. However, the preventative measures themselves are as complex as the original systems, making it impossible for a layperson to assess the risks in any specific scenario.

For example, commercial aviation is, statistically, incredibly safe. But, to evaluate how safe a particular flight is, you would need years of study and years of training to know all of the factors involved and assess how each of them affects the safety of the flight. The same can be said for other complex systems: A career professional or expert in the field can evaluate the risks in a given scenario, but the average individual cannot.

So, when navigating systems we do not understand, we tend to aggregate the intractable (but ever present) complexity of each interaction into a single statistical outcome, using publicly available data and conventional wisdom, for simplicity’s sake. Otherwise, it would be too difficult to keep up with the pace of modern life. In doing so, we also relinquish most of our agency over the situation. After all, there’s not much we can do if we don’t really know what’s going on.

Thus, our decisions begin to resemble steps on a Markov chain. In statistics, a Markov chain is a process where what comes next depends only on the current situation. Likewise, we allow the simplified statistical information that’s publicly available to substitute for true understanding of the systems we’re interacting with when assessing what will happen next. This may not strictly be true, mathematically speaking, but, outside of our own areas of expertise, we genuinely don’t have the ability to more deeply comprehend or influence the systems we’re interacting with. Will this salad contain E. coli? “Probably not.” Will this plane crash? “I heard commercial aviation is safer than driving.” Will that bridge collapse while I’m on it? “It should be safe, I’m sure they’re keeping an eye on it.”

The trouble with statistical models is that it can be tempting to use them where they don’t belong, out of convenience. When many of the decisions we make each day seem to be links on a Markov chain, we may view other choices where we actually can understand the situation and influence the outcome in the same, probabilistic way, out of habit. In doing so, we relinquish the pockets of agency we had between the great titans of complexity.

When we board Mr. Markov’s merry-go-round, every decision gets simplified into a yes or no question, and the answer is whatever seems likeliest to be correct. In an age where every question imaginable has already been asked online, it requires little effort and even less thought to aggregate the opinions of strangers into a probabilistic solution for our every uncertainty. This solution might have little to do with you as an individual or the specific context of your situation, but it will blend right in on the statistical carousel, if you allow it.

If you’re not sure if attending college is worth it, there are ROI stats for every school. If you’re choosing between different trade schools or universities, there are rankings along every imaginable dimension. Once it’s time to pick your major or your trade, you’ll be inundated with post-graduation employment stats. Then come the median salaries for every position at every company, the “Best Places to Work” rankings for every locality, the crime statistics for every neighborhood, the standardized test scores for every school district. The veracity and relevance of these stats are, at times, murky, but, aside from the occasional curiosity rabbit-hole, it rarely seems worthwhile to spend time dissecting this endless stream of information.

The decision to not simply be a passenger on the ride requires conscious effort. It means acknowledging that there are far more complex systems at work in our lives than we will ever have time to understand in a lifetime, but doing so without delegating our measures of free will to the marching stride of invisible giants. And, to be clear, we can do much more with our windows of agency than our ancestors could a millennium ago. Even a century ago, no one would be able to imagine the ocean of knowledge we now command at our fingertips, and the technology we can now purchase for a paycheck would seem like magic.

It takes balance to have one foot on the carousel, and one foot hanging off, ready to step off at the right moment, but prepared to hop back on when appropriate. It’s also a deeply personal practice; every individual has a unique set of skills and traits that qualifies them to take full control of their life under specific circumstances. And, as easy as it is to ride passively through life, it’s also possible to tip too far in the other direction, to believe you understand and control far more than you actually do, especially when you can always find armchair experts on the internet who agree with even the most outlandish of your opinions.

And so, to know when we should act, we can only keep honing in on who we are and what we are capable of; we can only keep reaching for, and falling short of, and learning about the things we desire, and why we desire them; and our footing at the edge of the merry-go-round will grow a touch more sure each day. And we shouldn’t be afraid of jumping a little too early, or of hanging on for a bit too long; the important thing is that we are able to decide for ourselves, and that we exercise that ability often enough not to forget it.

At least, that’s what I think. But what do I know?