This financial toolbox is modeled on the HASH Prisoner’s Dilemma simulation. This game theory simulation has agents using strategies to play a scored game with each other at each time-step, with the strategies which result in the higher scores being adopted by neighbouring agents. Our simulation operates similarly, except instead of being trapped in a Prisoner’s Dilemma, the game we’re playing is trading a portfolio of assets, and the score is the value of the portfolio.
We simulate a grid array of agents, each with their own portfolio of assets. At every time step, the agents make trades with their neighbours based on a trading strategy. If a neighbour’s strategy is successful, that is - it makes the neighbouring agent’s portfolio more valuable, it will be adopted by other agents. We can track the prevalence of strategies over time to see which ones are more successful.
We use agent-based modeling to model both the behavior of individual agents, and the system of trades as a whole. This allows us to understand which strategies fare the best, and we can also draw insights about systemic risk.