Why is the past so much easier to predict than the future?

Because it’s already happened.

In 2013, I personally made two very bad decisions that ended up causing a lot of problems. I had no way of knowing that they would be bad decisions because all of my analysis said they would be good ones. (In my defense, I did make good decisions last year as well).

Now, firmly seated in 2014, however, it is obvious that those two decisions were very bad. It is then quite easy to say that those decisions never should have been made because elements of history logically led to their inevitable outcome.

This is just as obvious as the Nazis winning World War II because in hindisight they had better equipment, a better trained army, and discipline. … but they didn’t win. And they didn’t win because of what now seems an equally “obvious” set of variables.

Dave Snowden and others call this “Retrospective Coherence”. Retrospective Coherence is, essentially, looking back in time and creating a coherent narrative around what happened. The accuracy of this narrative is always suspect because the narrative itself can only be based on assumptions of the motivations and methods of decision making that happened in rooms in which we were not present.

When I was a civil engineer, I needed to have a policy from Lloyds of London known as “errors and omissions insurance”. In essence, I was being insured for overlooking something that might cause a problem in the future. I was not being insured against negligence - I was being insured against accidentally not doing something right. I was being insured against Retrospective Coherence.

There are some people who define Retrospective Coherence as a scientific process of interpreting the past. There are others who define it as sensemaking. There are others who define it as fabrication. All would be correct.

The problem lies not in available data or in an unadulterated narrative - the problem lies in the known and unknown motivations of those looking backwards. Here we hit what interests me about RC: it’s a cumulative cognitive bias system.

There are too many to detail in a single blog post, but these cognitive biases not only frame how we initially interpret an evolving narrative, but they also impact how we interpret new data, how willing we are to be surprised by alternative explanations, and our ability to empathize with all the actors and decision makers involved. Further, they impact how seriously we treat the problem given the conclusion, what elements of the problem we personally fixate on, and the sample sizes of data we’re willing to accept.

The number of cognitive biases that impact Retrospective Coherence, and the fact that most people engage in it automatically, to me makes it what I call an Uber Bias. An Uber Bias is a cognitive bias that is made up of many other cognitive biases, but is describable as its own system. Retrospective Coherence, The Planning Fallacy, and maybe falling in love might be the three big Uber Biases for me at the moment.

This isn’t to say we shouldn’t make sense of the past, plan for the future, or fall in love. It merely is saying that they are all acts of surprising complexity.