We present a new click model for processing click logs and predicting relevance and appeal for query–document pairs in search re-sults. Our model is a simplified version of the task-centric click model but outperforms it in an experimental comparison.