I found it very interesting how Ned Hall tried to draw a distinction between what he sees as the two definitions of a prediction. Namely, he contends that the basic definition of prediction is merely a forecast of what will happen in the future and that a second definition is where we try to base intervention strategy on predictions of what happens depending certain variables. While this delineation of definitions on face value seems acceptable, I would argue that these two are essentially the same definitions. The latter definition is simply a more complex version of the former. Indeed, the first definition, in more detail, is a forecast of what will happen based on whatever we think current variables will be however far into the future. On a fundamental basis, that draws upon the same framework as the latter definition as we are basing our forecast based on where we think certain variables will be. This is especially pertinent to the creation of machine learning algorithms and how we utilize them in e.g., trading securities.
Useful citation: Chen, Guanting, et al. “Application of Deep Learning to Algorithmic Trading - CS229.” Stanford University, cs229.stanford.edu/proj2017/final-reports/5241098.pdf.