Representing Stages and Levels of Automation on a Decision Ladder

We propose that representing stages and levels of automation on a decision ladder (DL) could help to identify information requirements for designing automation interfaces. We look at automated financial trading systems, a domain with variable degrees of automation (DOA). We give examples of modelling a financial trading task for two DOAs: basket trading (a low DOA) and trend following trading (a high DOA). On the resulting DLs, both human and automated information-processing activities are presented. The steps and states of knowledge allocated to automation are first categorized by the commonly known four stages of automation, and then shaded to represent the level of automation in each stage. This work advances the understanding of automated trading, and automation in general, and may provide a deeper representation of human-automation interactions and thus better understanding of design requirements.