Paper short abstract:

This paper examines a series of historical artefacts used to predict the future from numbers. It looks at them as designed objects—products of a specific time, place and set of intentions—as well as steps in the trajectory leading up to the present forms of computational predictions.

Paper long abstract:

This paper examines the history of statistical prediction through the lens of designed artefacts; the tables, graphs, tools and models which have enabled and communicated predictions since the rise of statistical thinking. It aims to put the current promises and fears around data-science into perspective by retracing the trajectory that led to today's systems of computational prediction.

Looking back into recent history reveals that contemporary hypes surrounding the increasing role of computational statistics in society are not solely caused or enabled by new technologies. They are new instances of systems created around similar motivations and politics. Contemporary concerns about black box algorithms for example, find a strong echo in 1902's USA and the Medical Information Bureau; a secretive system which standardised and codified medical impairments in a booming life insurance industry.

Coming in from design, particular attention will be given to the aesthetics of statistical forecasts to visually decrypt the narratives surrounding prediction systems. We will look at how the documents, formulas or devices used in these systems have embodied an aesthetic of accuracy which, after two centuries of evolutions and refinements, is ubiquitous in today's technological and scientific landscape. We will examine the role these aesthetics have played in shifting the tensions between mystical notions of predicting the future on the one hand and verifiable, quantified modes of prediction on the other. These tensions are still very present in today's computational prediction systems, for example in the perceived powers of 'code'.