Our brains love dopamine. We get a hit of adrenaline and dopamine when we’re “right”. This addiction to being right works well in our lightning fast world where decisions are expected to have been made 2 weeks ago. We quickly get to “right”, get our hit of adrenaline and dopamine, and move on to the next decision. It also worked well when we were being chased by saber-toothed tigers… run fast … now, turn left or right? I turned left last time and lived (giving me a hit of dopamine and adrenaline). Quickly now … OK, I’ll turn left.
In order to arrive at “right”, we tend to use the tried and true method of calculating the worst and best case scenario. Let’s unpack the typical process of this method when making an investment decision:
- Derive a formula that delivers a single answer (we must be either right or wrong after all. How can we contradict our decades of formal education?)
- Decide what is known about every variable in that formula with a bias towards having more knowns that unknowns (how can we be exactly right with a bunch of pesky unknowns?)
- Drop known values into all of the “known” variables
- Drop the worst possible value into each of the unknown variables
- Run the formula to get the Worst Case
- Go back to step 4 and repeat for the best possible value
- Run the formula to get the Best Case
Pretty simple. I’m sure we’ve all done this a million times. It works… until it doesn’t.
The major issue with this method is that it generates 2 decision points that are extremely unlikely to happen. You could call them Black Swans to use the Nassim Nicholas Talib vernacular. In pure statistical terms the worst case and best case have a less than 0.3% chance of happening. It’s like trying to decide to go skating outdoors tomorrow by only considering that it could either be too hot because the earth will crash into the sun or that it will be too cold because the next ice age will have started. Crazy, right?
This post is not about how likely it may be that your worst case scenario is truly bad or that your best case is truly good. That is fodder for another post. What I hope to show is that this simple method should *only* be used when it fits the question you want answered.
Let’s examine a common question in the realm of risky investment decision making. In case you’re wondering about the use of the word “risky” here, don’t all investment decisions carry risk in the absence of a crystal ball?
Question: “I want to invest, but will I lose more than $x? I can’t afford that.”
There are some who believe that you shouldn’t be making risky investments if you can’t afford to lose it all. That’s always rung a bit hollow for me as it depends on the, very relative, definitions of risk and affordability. What is risky for you might not be so for me. Same with affordability. That means there is no such thing as an investment that is absolutely too unaffordable. It all depends on your definitions. Therefore, there is room to define risk. That is, undoubtedly, where worst case definition and decision making comes in.
Let’s say I asked you to invest $10k in a new restaurant I want to open. You have the $10k and you can afford to lose half of it. Those of you in the “you can’t be an investor and not expect to lose it all” camp will be bristling at this point as we all know how risky new restaurants are. Let’s put that aside for a moment and focus on how I get comfortable making the investment.
How do I discover what my worst case scenario is and if it’s better or worse than a loss of $5,000? Any number of ways including researching similar restaurants that have failed and how much that failure cost its investors, and asking experienced restaurateurs their opinion on the worst and best case. Let’s say I get an answer from both those sources and it’s a max of a $3,000 loss. With a sigh of relief, I hand over my $10,000 and wait. Then I start to wonder … When might I see the return? How likely is it? It’s this last question that is the kicker. When you did your research, you determined $3,000 as the worst case. That means there is no outcome that could be worse. Surely you don’t believe that given it’s a restaurant, but how much worse could it get and how likely is it that worse worst case could happen?
THAT is the problem with predicting that the future will occur at a specific point. No one can predict that. I don’t care how good or experienced you think you are. The only reasonable way to accurately predict the future is by articulating it as a range of possible outcomes with a likelihood of occurrence for that range. Articulating the future at a single point is just plain lying. Ok, so this last bit is a blog post that I will be writing. Stay tuned to this channel!
Let’s go back to our restaurant investment example. If a $3,000 loss isn’t the worst possible outcome, what is? That’s fairly obvious, it’s $10,000. At the risk of agreeing with my “you can’t be an investor and not expect to lose it all” friends, there’s a better way to decide to invest than having the means to lose all of your investment every time. What if you could articulate your actual risk as a 40% chance of losing between $2,000 and $3,000? Or an 80% chance of a -$2,000 to $1,500 return?
This is the only way to express a future event. Remember, no matter how confident we make ourselves with the research we do, the people we talk to or our own opinion, the future will remain decidedly uncertain.
So what are the worst cases when you decide using worst/best case “analysis”? For starters:
- You could underestimate the worst case and lose way more than you thought possible.
- You could overestimate the best case and make nowhere near as much as you thought possible which would have made an investment in something much less risky a better choice.
How bad these outcomes could get is a function of the magnitudes of those outcomes above and other unexpected and unwelcome outcomes not mentioned.
I was recently reminded of the value of thinking in probabilistic terms by my esteemed colleague Dan Toma (if you don’t know Dan, you owe it to yourself to check out his blog and book The Corporate Startup). Dan came to me with a problem that he had estimated with a worst and best case scenario. He did so because he knew that these 2 end points were insufficient to make an accurate decision. We modelled the problem using a Monte Carlo based method I use frequently that I call The Unison Method (something else that will feature in the aforementioned future post). What emerged were ranges of outcomes and their probabilities that surprised us. We were surprised because our expectations were anchored on the best and worst cases. What we didn’t initially consider was the extreme unlikelihood of those 2 outcomes.
How do you avoid anchoring yourself with unlikely to occur Black Swans when making decisions? I’d love to hear from you.