Parameterised response zero intelligence traders

Abstract I introduce parameterised response zero intelligence (PRZI), a new form of (ZI) trader intended for use in simulation studies the dynamics continuous double auction markets. Like Gode and Sunder’s classic ZIC trader, PRZI generates quote prices from random distribution over some specified domain discretely valued allowable prices. Unlike ZIC, which uses uniform to generate prices, probability is such way that its mass function (PMF) determined by real-valued control variable s range $$[-1.0, +1.0]$$ [ - 1.0 , + ] determines strategy trader. When $$s=0$$ s = 0 , behaves identically strategy, with PMF; but when $$s \approx \pm 1$$ ≈ ± 1 trader’s PMF becomes maximally skewed one extreme or other price range, thereby making it more less “urgent” generates, biasing towards away limit price. To explore co-evolutionary populations traders dynamically adapt their strategies, show initial results long-term market experiments each simple stochastic hill-climber algorithm repeatedly evaluate alternative -values choose most profitable at any given time. In these profitability particular -value may be non-stationary because time can depend on mix strategies being played time, are themselves continuously adapting. Results demonstrate population traders’ exhibit rich dynamics, periods stability lasting hundreds thousands interactions interspersed occasional change. Python source code traders, hill-climber, have been made publicly available GitHub.