Opinion: the use of prediction marketplaces could help us all to make sounder decisions
By Paul Cuffe, UCD
The last two years of on-again/off-again lockdown restrictions often created agonising decisions for many of us. For instance, you may have spent months wondering whether you needed to reschedule your wedding. Right now, the Covid situation seems optimistic, but we might still wonder if there are new strains and further lockdowns waiting for us down the road.
While there was no ideal way to navigate the last two frustrating years, it would have been helpful to have numbers to put next to our uncertainties. If the weather forecast tells me there's a 10% chance of rain tomorrow, I might leave the umbrella at home, but a 80% chance means I’ll be sure to dress for a downpour! Wouldn’t it be insightful to have a similarly specific forecast on the prospects for a shooting war in Ukraine, or the emergence of a new and unwelcome Covid strain? US president Joe Biden has offered his "guess" that Russia will move into Ukraine, but just how strong of a prediction is meant by such language?
One handy way to crowdsource a precise forecast is to set up a predictions marketplace. Suppose you are in charge of an Irish business with significant operations in Ukraine: how nervous should you be, do you need to vacate the country altogether or should you sit tight for now?
A large body of evidence shows that such consensus beliefs are uncannily accurate and that the wisdom of the crowds really works.
First, define a clear yes/no question on this point: will it be possible for Irish citizens to freely travel to Kiev on 15th August 2022? Then, using an online platform built for the purpose, make special tokens available which can be cashed in for €1 apiece if travel does turn out to be permitted on that date, but are worthless otherwise. Then, sit back and let punters trade these tokens amongst themselves in the run-up the fateful day.
You can learn a lot by watching the prices at which these tokens change hands. If Russia starts making peaceful noises, for instance, punters will want to stock up on these token in the hope of securing a future windfall, and will bid up their price accordingly. On the other hand, if the sabre-rattling grows louder, traders will want to offload these suddenly-less-promising tokens, and this sell-off will force prices downward.
A useful side effect of all this trading push-and-pull is that the token’s price is actually a high quality forecast of how likely the future event is to occur. If one day the going rate for these tokens is €0.64, this means that the punters collectively believe that there’s a 64% chance that you’ll be allowed to conduct your business in Kiev come mid-August.
There is simply no way of knowing definitively when a fresh strain of Covid might emerge, or what the Russians might do next
Remarkably, a large body of evidence shows that such consensus beliefs are uncannily accurate and that the wisdom of the crowds really works. Crucially, players in a predictions marketplace have skin in the game: bad trades cost them real money. The self-appointed military expert on Twitter can spout false predictions in a snappy 280 characters at no cost to themselves, but a punter pays a heavy tax for being wrong. Prediction marketplaces reward only the accuracy of a belief, regardless of how unfashionable or politically inconvenient this may be. This is why they provide solid forecasts in a way that pundits, panels and insiders cannot.
Not many of us are in charge of international businesses, but we’ve all had frustrating moments over the last two years when a solid forecast on the easing or tightening of lockdown restrictions would have been very useful. Unfortunately, there are few active prediction marketplaces on these types of question, so we had to rely on our own intuitions and guesswork to navigate all those tricky lockdown uncertainties. Our collective forecasting expertise is instead deployed on such matters as predicting the swiftest horse in the 2.45 at Cheltenham. This seems a shame: we are enduring all the real social harms of gambling, but we aren’t even capturing its one potential benefit, which is creating useful forecasts of how an uncertain future might unfold.
In my own area of electrical power engineering, it is becoming very important to forecast the amount of wind power that will be available on the national grid tomorrow, or the day after, or even into next week. With my doctoral student Mahdieh Shamsi, we are exploring how predictions markets can provide such forecasts, hopefully making the national grid more secure while also making wind farms more profitable. We’re also exploring how the technology that underpins cryptocurrencies (like Bitcoin) could be adapted to host online prediction marketplaces in an automated and transparent way.
Humanity will never invent a perfect crystal ball. There is simply no way of knowing definitively when a fresh strain of Covid might emerge, or what the Russians might do next in Eastern Europe. Any long-term decision entails a little risk, whether that's booking a flight for next summer, rescheduling your wedding, or even expanding your business into a politically volatile country.
Prediction markets can’t remove these risks, but they can at least tame them a little, by providing clear numerical probabilities that represent a credible consensus forecast. Sound decisions (personal, political or otherwise) shouldn’t be based on wistful thinking nor blind pessimism: prediction markets can offer us some firmer footing for navigating an uncertain future.
Dr Paul Cuffe is an assistant professor in the School of Electrical & Electronic Engineering, and a member of the Energy Institute at UCD
The views expressed here are those of the author and do not represent or reflect the views of RTÉ
