With election season almost on us on both sides of the Atlantic. Who do you think will win?
“I’m doubtful Donald will do it again”.
“There is a good chance Kier is going to be our next PM”.
Voters go to the polls in Russia too, are Putin’s chances of a fifth term “certain”?
Everyone has an opinion. From big name television pundits to groups of mates at the pub. Views everywhere. It’s only going to get worse as campaigns ramp up. Good luck with that slightly sozzled uncle around the Christmas table.
On 28 October boxing fans were glued to their screens for Tyson Fury’s bout against Francis Ngannou.
Fury undefeated in 35 fights. WBC and lineal heavyweight champion. Ngannou a man who had never boxed professionally before. A mismatch on paper.
According to experts the underdog didn’t “have a chance”. It would be a “comfortable” win for the Brit. Tyson in “a few rounds”. Only MMA fighter Tom Aspinall gave his fellow octagon specialist “a chance”.
On the night it was close. Split decision. The Gypsy King knocked down in the third round. Squeaky-bum time for a man due a unification fight with a vastly more experienced opponent in December.
Perhaps it’s a good time to remind ourselves of the fallibility of this sort of crystal ball gazing.
UK civil servants know that ministers are increasingly demanding data-led decision making. Predictive data has never been in bigger demand.
It’s not enough to write a policy paper, or undertake a futures analysis, with a purely qualitative approach.
At universities in the UK politics students cut their teeth on a number of important articles. Anatomy of a Failure: The Decision to Land at the Bay of Pigs is one that sticks in the mind.
Vandenbroucke breaks down the key crunch points that led to one of the biggest foreign policy failures of the Cold War.
U.S. Joint Chiefs of Staff (JCS) told JFK the plans had a “fair chance” of success.
What would that mean to you?
Kennedy was content, he gave the plans the green light.
But what did military chiefs actually mean? 3 to 1 against.
Brigadier General Gray, who published an analysis of the CIA’s plans before the invasion, didn’t use any numbers in his report to JCS.
Lots of research on the drawbacks of verbal probabilities has been done since in every field imaginable. From medical diagnosis to stock advice.
The pitfalls of are nothing new. So why do we keep doing it?
One reason is because busy decision makers find too many predictions dazzling. Barack Obama, when faced with multiple probabilities of between 30% and 95% that Osama bin Laden was in Abbottabad, reverted to a simple summary. “This is 50:50%”.
Another is that making good numerical predictions can take time. Debiasing. Finding the right forecasters. Building a good team. Aggregation. Setting the right question. Calibration. When decision makers need to make a quick call they often revert to type.
A third is that too many civil servants and politicians have a background in humanities and arts-based subjects. Using numbers feels uncomfortable.
Why do people prefer words? Studies point to all sorts of reasons. It feels more natural. More familiar. Easier. Intuitive. Pragmatic.
But research from the University of Essex suggests it’s a bit more complicated. It depends on the type of uncertainty you’re dealing with.
When it comes to distributional uncertainty people might prefer numbers. This is the sort of uncertainty that has a clear set of alternatives in people’s heads (“the chance of the coin landing on heads”). When the question is more vague, and relies on the people’s own knowledge or personal disposition, words are the order of the day.
It depends on the context too. Research shows that patients prefer to have numerical estimates attached to a prognosis from a doctor.
There is room for optimism for those of us who feel numerical probabilities should be prioritised where possible.
A meta analysis of 21 studies by a team of German researchers shows there is not as much disagreement as you might think in the academic community about how verbal probabilities should translate to probability scores.
Practically speaking, putting percentages next to terms like “roughly even” or “very likely” can reduce confusion. Even if it means using something as simple as a guideline table, brackets, or a ‘mouse over’ tooltip.
There are a limited number of events for which a precise probability can be computed with 100% accuracy. Leaders might ask: why bother?
Our partners at Good Judgment Inc. are experts in putting together expert forecasts about the future quickly. Iterating and revising numbers when the goalposts change.
ForgeFront’s Delivery.Ctrl methodology translates this into actionable present-day policy and strategy implementation. Making it accessible for busy decision makers with the latest digital tools.
As we look to 2024 the world is on the tipping point for a huge number of transformational societal, economic, technological, environmental and geopolitical changes. For this boxing fan, we could even be getting the first unified heavyweight champion in two decades. Getting this right has never been more important.