Models splinter when you look at them closely – Part 2

In the last post Models splinter when you look at them closely – Part 1, I introduced Emmanuel Derman’s book Models. Behaving. Badly. and discussed a few different types of models with the key idea that we use models to emphasize some features that are important, and eliminate others that are not important, for a specific purpose.

In this post I’ll cover a few aspects of models and their limits, and conclude in part 3 with benefits and risks.

Models leave things out

All models are metaphors – our model is like the thing being modeled, but it isn’t exactly the same. The question is whether the things that we are ignoring are important for the purpose of the model. In the case of the fashion model, trained buyers can make adjustments based on their knowledge of clothing, store displays, and customers.

A good architect knows enough about buildings to know what is not represented in the tabletop scale model. This is not always the case, as there are many stories of redesign, delay, and cost overrun due to visions that are difficult to execute. This is a key point. We spoke of models eliminating unnecessary aspects for a specific purpose. If the purpose is deciding on the shape of the building and how it interacts visually with a surrounding environment, a tabletop model may do the trick. If the purpose is how to effectively build a structure, or how it will interact physically, chemically, or aerodynamically, or how it will feel to people who work or live or shop in it, a tabletop model will fall short. Virtual reality, computer-aided drafting and design, and other modeling techniques are required.

Models overlap

We use layers of theories and models that are connected (or not) through observation and measurement – analytic continuation, Derman calls it. No model includes everything.

In the physical world we have overlapping theories that span from the subatomic realm to the farthest reaches of the cosmos. (There are gaps at the extremes that are patched with what are to me unsatisfactory solutions i.e., dark matter.)

On a simpler scale, when I lead an outdoor trip, I use a road map to get me to the trail head. I have a trip description that gives me a high-level orientation to the trip. I use trail books, maps, satellite images, weather forecasts, and other resources to decide when and where to go and to develop a detailed route plan. And I’ll layer GPS and topographical map data for terrain and exact positioning. But when I get there, none of these representations shows that:

  • the trail has been washed out
  • signs or markers are missing or wrong
  • someone came on the trip unprepared
  • the weather has changed from 28 and sunny to 12 and pouring rain.

There are limits to all models – the important part is to know what the limits are, to establish measurements to identify when limits are being reached, and to use new methods when the boundaries are exceeded.

In organizations it works the same way. Our sales information technology connects to an accounting information technology. Each information technology is designed with a user interface that connects with the people we’ve hired through our human resource processes and systems. We pay our people in a way that enables them to perform and makes them want to stay. Each of these areas has a body of knowledge, a best practice, and underlying qualitative model.

(What we lack in most cases are quantitative computational models that show us how these systems, processes, and people could interact over time to affect long-term performance. It can be done through system dynamics modeling and other methods – this is the leading edge that I’m interested in – ask me!)

Let’s leave it there for now and conclude with the summary: there are different types of models that we use for different purposes, and in general we are emphasizing some features and eliminating others that are not important for a specific purpose.

See you next time in the final part.

Interested in models or have comments? Let me know at info@thinkitation.ca.

And sign up for the newsletter!

Disclosure: I have no personal or business relationship with Derman or his book.

Reference: Derman, Emmanuel. Models. Behaving. Badly. Why Confusing Illusion with Reality Can Lead to Disaster, on Wall Street and in Life. New York. Simon and Schuster Free Press. 2012

By |2018-01-29T00:51:07+00:00January 23rd, 2018|Categories: Any Sector, Modeling|

About the Author: