The Keystroke-Level Model is a simple model for predicting the time it takes an expert user to perform a task on an interactive computer system. It is based on counting keystrokes and other low-level operations, including the user's mental preparations and the system's responses. The model uses heuristic rules to predict where mental preparations occur and has been tested against data on 10 different systems, with a prediction error of 21 percent for individual tasks. The model is compared to simpler versions and is discussed in terms of its potential role in system design. The model addresses only one aspect of performance: the time it takes expert users to perform routine tasks. The model assumes that the time for an expert to do a task on an interactive system is determined by the time it takes to do the keystrokes. The model includes four physical-motor operators (K, P, H, D) and one mental operator (M), plus a response operator (R). The model has been validated through an experiment involving 1,280 user-system-task interactions, with results showing that the model's prediction error is 21 percent. The model has been used to predict task times for various systems and tasks, and its accuracy is discussed in terms of its ability to predict execution times. The model has also been used to illustrate how it can be applied in practice, including the use of parametric analysis and sensitivity analysis to examine how changes in task or model parameters affect predictions. The model has been shown to be effective in predicting task times for a wide range of user-computer interactions.The Keystroke-Level Model is a simple model for predicting the time it takes an expert user to perform a task on an interactive computer system. It is based on counting keystrokes and other low-level operations, including the user's mental preparations and the system's responses. The model uses heuristic rules to predict where mental preparations occur and has been tested against data on 10 different systems, with a prediction error of 21 percent for individual tasks. The model is compared to simpler versions and is discussed in terms of its potential role in system design. The model addresses only one aspect of performance: the time it takes expert users to perform routine tasks. The model assumes that the time for an expert to do a task on an interactive system is determined by the time it takes to do the keystrokes. The model includes four physical-motor operators (K, P, H, D) and one mental operator (M), plus a response operator (R). The model has been validated through an experiment involving 1,280 user-system-task interactions, with results showing that the model's prediction error is 21 percent. The model has been used to predict task times for various systems and tasks, and its accuracy is discussed in terms of its ability to predict execution times. The model has also been used to illustrate how it can be applied in practice, including the use of parametric analysis and sensitivity analysis to examine how changes in task or model parameters affect predictions. The model has been shown to be effective in predicting task times for a wide range of user-computer interactions.