Learning from the past to help decision making today

Our previous newsletter looked at lessons from aerospace decision-making as applied to other endeavours, for example when investing in the markets.

Specifically, we looked at how time, and our perceptions of time, can influence our decision making.?Understanding our perception of time and our time horizons can give us increased clarity in a 24-hour world of noise.

Let's look at the following three charts: the first is over three months and shows how the markets appear to us today. The second expands out to three years. The last shows the absolute truth over the previous twenty years. The markets have seemed chaotic and opaque over the last six months, mainly the last three. Commentators and investors alike have struggled to find sense in a conflicting story of supply-induced inflation and signals of an impending recession.

If one were to decide or even attempt to plan within this narrow time frame, the outcome would likely be suboptimal. Indeed the green line of the VIX Index shows the enhanced volatility of the period. It's hard to see a pattern let alone find the information on which to make a decision (apart from waiting for time to pass).?

Looking at the data from three years, the trends become somewhat clearer, we can see the impact of the imposition of the Covid lockdowns with the spike in the VIX in March 2020. The artificial stimulus, both monetary and fiscal drives the S&P higher, and the blue line of the 10yr Treasury started to respond to the inflationary impulses driven largely by the war in Ukraine.?

?Looking at the twenty-year picture and beyond (in the bottom graph), one can distinguish specific events, trends and themes.?There are two clear spikes in the VIX. one driven by the 2008 financial crisis, and the second major one driven by Covid (yellow arrows).?

Indeed when one looks at the twenty-year graph, it's easy to make out the excess valuation bulge above the red trend line of growth within the S&P. In this time frame, the recent sell-off makes sense within that reality. One can also make out a longer-term trend (represented by a pink arrow) in the decline of Treasury yields which predates 2008 and then exhibits a so-called technical 'double top' between 2012 and 2018.?

So, when we look at decisions within a temporal context, pictures such as this remind us that decisions are sometimes part of a predetermined process, occasionally immediate, and sometimes part of an adaptive and reactive process. But they are always best driven by an awareness of the background picture.?

As we mentioned before, we are all too familiar with these phenomena in aerospace. We have learned to evolve our processes to give the decision-makers the most accurate picture of reality.

There are linear pathways at work, with diagnosis, options, choices, and a plan. What comes as a critical point in that pathway is a decision point. The decision point can come before or as the fundamental part of a plan. Indeed, after a decision, there is a general need to allocate tasks to bring the desired outcome into reality.

?In her seminal work on Ai decision making,?Link,?Lorien Pratt describes a decision as:

?"A thought process that leads to actions, that lead to outcomes."?

?It's an excellent definition that leads us to some interesting points worth breaking down.

?Process?

Sometimes predetermined, sometimes immediate and reactionary. In aerospace, a key element to its success are the learning processes that have evolved from learning from history. Mistakes are shared, lessons passed on and there is a semi-religious reliance on 'No Blame' cultures.?

Actions?

This is where we get to it, ideally as part of a plan, but sometimes as part of a predetermined order. In algorithmic trading, there will be buy and sell signals, based on modelling that attempts to secure a particular outcome. In one word, a?plan.?

Outcomes?

Interestingly, a large national airline has recently stated in its training material to pilots, that there is a common element missing when decision making is studied, Luck. Luck is the special ingredient in all outcomes, no matter how hard you plan, prepare and train, luck is often a significant element. The trick is to try and reduce the amount of luck needed for a successful outcome. Can you teach Ai luck?

Sounds simple. But as the autonomous vehicle industry has discovered, even millisecond decisions can have infinite variables. That's why it's proving to be such a complex task.

There are various layers of decision-making automation, and one of the most interesting is?Decision Intelligence.?We'll be looking at that next week, see you all then,?

As always we'd like to thank Sentieo for their data research services, you can find them at:

www.sentieo.com?

This is an opinion piece, written by the authors and does not constitute investment advice.