Core Concepts #6: Dynamics
a.k.a. “Everything won’t stop changing”
This is part six of a seven-part series of posts on the core concepts of tradespace exploration, designed to help beginners become familiar with terminology and the general structure of tradespace data. Click here to find the other posts in this series, as well as other Tradespace 101 posts.
Relatively few decisions result in value that is delivered in a single burst — most of the time we gain value over time as we continue to reap the benefits of our chosen alternative. This introduces the possibility of dynamics to the problem. After all: things change over time, and the value of our decisions is no exception. When value is accrued over time, we sometimes refer to it as a value trajectory, which can go up or down depending on the changes that occur.
A quick note before we go any further: value can be dynamic even if the future is completely certain, as long as it is accumulated over time. For example, taking a round trip vacation to the equator, the North Pole, and home again results in a decision on what clothes to wear that must account for the dynamics in comfort (value) caused by temperature change. And yet it is certain that the equator is hot and the pole is cold! Our value trajectory will change dramatically from location to location for any given outfit, and we can accurately predict that trajectory.
Despite that, uncertainty is by far the most common driver of dynamics: epoch variables are outside of stakeholder control and have a tendency to change repeatedly over time. When dealing with value trajectories, we must necessarily consider the future, and the future is (almost) always uncertain. Very few stakeholders will be fortunate enough to know exactly how the surrounding context and/or their needs will change when making decisions with effects lasting years, months, or even days.
How do we capture value trajectories using a tradespace? If we have set up for Epoch-Era Analysis, this is when we begin to utilize eras: sequences of epochs that represent the epoch variables (uncertainty) changing over time. Each era tells a story of how uncertainty could impact the value of a given alternative over time — we can follow along as it rises or falls or holds steady, and we can calculate metrics based on the lifetime of the alternative (e.g., accumulated value).
Epoch variables will often change on different timescales: i.e. some on the order of hours, some on weeks, some on years. Eras should account for these differences in order to best represent realistic potential lifetimes for the alternatives. The goal is to divide the lifetime into blocks of time that are aligned with the natural rate of change of uncertainty. In practice, that means the epochs in an era are not required to be the same duration: we don’t simply discretize time into one-day chunks and then force all epoch variables to change at the same rate. Discretizing the lifetime only when something actually changes enables our analysis to directly address the dynamic value of the alternatives.
Generally, we create eras in one of two ways:
- Narrative. The narrative approach is similar to scenario planning and “what if” analysis, in that we manually construct an era in order to demonstrate and explore a specific potential lifetime. We might explore a “normal” era, where uncertainty changes as we expect it to, as well as best-case/worst-case eras or high-variability/low-variability eras. This is fast, easy to understand, and useful for developing insight into which alternatives are most affected by which uncertainties.
- Sampling. We can use a computer to sample large quantities of eras and then calculate statistics for each alternative’s lifetime metrics: e.g. average accumulated value or variance of resources spent. This is mathematically powerful but more time consuming, and — for some types of statistics — requires knowledge of the probability distributions associated with the epoch variables in order to create “rules” that are used to generate a realistic sample.
Keep in mind that understanding dynamics is often more complicated than just comparing averages. The success of many decisions is path-dependent, meaning the order in which uncertainty unfolds can matter just as much as the epochs themselves. For example, imagine you were considering opening a new business. If you knew that 9 of the next 10 years would feature a strong economy but the 1 other year would be a recession, you might say that on average that sounds pretty good! But consider the difference between eras where that recession hits in year one versus year ten.
- In year ten, with nine good years under your belt, your business has probably generated a substantial loyal customer base and maybe even built up a small war chest. You can probably weather a yearlong recession and make it out fine.
- If that recession comes right after you open, you have a big problem. Maybe you needed to take out a loan to get started and now you can’t make your payments, because customers are staying away from new stores. Your business might fail before you can get off the ground!
Wrinkles like this are what make it so important to consider uncertainty not just as an abstract set of probabilities but also as a complex and dynamic challenge. Stakeholders that ignore these dynamics do so at their own peril!
We have mentioned dynamic value and dynamic uncertainty, but alternatives can also be dynamic. Most decisions consider only static alternatives: one alternative is chosen, and that alternative must be able to deliver value through the subsequent dynamics it is exposed to. However, sometimes it is possible to change the alternative over time and, unsurprisingly, dynamic alternatives are often particularly well-suited to maintaining value in the face of dynamic uncertainty. Usually this requires additional resources, but when uncertainty is high it can be well worth the cost — especially if no single alternative is passively robust enough to perform well across the lifetime. Dynamic alternatives are often described as having flexibility or changeability, such as a convertible that can raise or lower its top based on the weather for maximum enjoyment. We’ll talk more about those in the next post!
Stakeholders dealing with a dynamic problem must also reckon with time preferences: relative preferences on when value is delivered. The most common approach to dealing with time preferences is discounting — the approach borrowed from finance that makes value gained now better than an equal amount of value gained later — especially if there is an opportunity cost associated with waiting. However, discounting is not appropriate for all problems and has been found to often undervalue the quality of the future. Adding to this challenge is the fact that many stakeholders do not even have a firm grasp on what their time preferences are, because they don’t regularly consider it when making decisions. Identifying key windows of opportunity is one method of working with a stakeholder to define the timeframe where the decision should have the most leverage/impact and thus (ideally) deliver the most value.
