This article examines a government attempt to impose testing of 4-year-olds as a baseline against which to "hold primary schools accountable" for children's subsequent progress. It examines the various forms of baseline testing in this experiment and analyses the misleading claims made for the "predictive validity" of baseline scores. The article also takes a broader look at standardised ways of tracking children's attainment and progress to the end of primary school and tacit assumptions of linear progress underpinning large-scale data-based accountability processes.