The following article was written for PAXsims by Ben Taylor (Defence Research and Development Canada) and Benjamin Williams (Professeur des Universités, IAE & CleRMa, Université Clermont Auvergne). The views expressed are those of the authors and do not represent the official policy or position of any agency, organization, employer or company.
The authors met through a workshop on Wargaming the Pandemic hosted by the King’s Wargaming Network that was held 1-2 April 2020. BW gave a presentation in which he set out an idea for a matrix game on the COVID-19 crisis that could be supported by quantitative epidemiological and economic models. BT had previous experience with matrix games and offered to collaborate on the idea. This project is therefore itself a product of the COVID-19 crisis as the authors are unlikely to have met or to have found a common project to work on without it.
We decided from the outset that we wanted to design a game that tackled the COVID-19 crisis in a country from the point after the initial lock-down measures had flattened the curve. This phase would require a balancing act by political leaders as they face challenges on three axes: economic, social and healthcare. We termed these the three frontlines of the battle against COVID-19. Our aim was for a game that would sensitise decision-makers to issues that they might face and one in which choices would be constrained by the cross-coupling between the frontlines; for example that returning people to work in offices would likely increase the rate of infection, or that a renewed lock-down would lead to public discontent. We also wanted to introduce some quantitative models to help elaborate upon the consequences of player actions.
We also decided that we did not want to build a detailed game around a specific country. Rather we wanted a tool that could be customised to any country. That required the game to have a generic framework to which national specific details could be added. For development purposes we settled upon the fictitious country of Bretonia which has a government structure like Canada and the economy of France. Our generic framework envisaged four players to represent key elements of the country; the national government, the lower tier governments, the business sector and the public health system. A fifth player, termed “The Crisis”, represents all other domestic groups, external actors and anything else that could happen to challenge the other players’ efforts. An example of the customisation necessary comes from different national approaches to healthcare funding. In Canada healthcare is a provincial responsibility, whereas in France it is mainly funded by the national government through the social security system. This difference would have to be represented in the roles and responsibilities of the two government players.
One of the first steps in designing the game was to develop an influence diagram that showed how various parts of the economy, business, government finances, social attitudes, the healthcare system and the pandemic itself are connected. This provided the reassurance that everything that we wanted to be in scope was captured. The model also provided insight to where knock-on effects (positive or negative) might be felt, which would provide for consistent adjudication.
We also built a dashboard that displays selected metrics grouped across the three front lines, a macroeconomic model, a model of the infection and fatalities and a slide deck for displaying new stories each turn. This latter part of the game was developed to provide some humour, some cultural flavour and to allow attention to be drawn to specific sectors of the economy. We also prepared a number of bad news stories to be injected if any of the economic or social metrics approached worrying levels.
Many design issues common to matrix games apply equally to this game. Among those that we encountered are:
The advantages of having players who have played matrix games before.
The need for subject matter experts to support adjudication if the results are to be realistic.
The challenges for players to switch between role-playing and becoming engaged participants in adjudicating arguments.
Whether the players should be left to solve the basic problem of opening the economy without triggering a spike in infections, or to subject them to additional external challenges, and in the latter whether it is best to script the injects or to have them occur randomly (the answer of course is “it depends”).
The balancing act between allowing players to discuss the proposed actions in detail and curtailing discussion in order to speed up the game.
The game has been run twice with participants from Europe and Canada using a video conference link with supporting text chat facility, a Google slides deck to share news stories, and Google sheets to share the dashboard of metrics and to provide an online tool to capture the participants’ assessments of the likelihood of success of proposed actions. This setup worked very well and participants felt that they could communicate with each other and access the information that was required. There was agreement that the game largely felt right, but that play was slow. The supporting quantitative models were not used extensively. In particular the epidemiological model implemented according to formulation drawn from the literature produced counter-intuitive results and proved impossible to fit to the observed progress of the outbreak in Canada. This placed a particular burden upon the adjudicator to determine how to adjust the dashboard in response to player actions.
Our next objective will be to design a discussion-based game without the matrix structure in order to compare the utility of the two gaming techniques in addressing the management of the COVID-19 crisis.