High redshift galaxy protoclusters are the precursors of today’s massive clusters; the sites of formation of the most massive galaxies in the present-day Universe. Using a sample of 37 clusters and protoclusters at 3.2>z>1.3, I will present the formation history of massive galaxies. I will demonstrate that these (proto)clusters form an evolutionary sequence, with the lower redshift (proto)clusters having similar properties to the descendants of those at high redshift. The colours of the protocluster galaxies deviate at z>1.6 from the passively evolving models commonly used to model cluster galaxies at lower redshifts. I will show that the observed colours of massive cluster galaxies are more consistent with cluster galaxies forming according to the cosmic star formation rate density of the Universe, with a truncation in star formation around z~1-2. Furthermore, I will show that the star formation in these massive galaxies must have been rapidly terminated to produce the observed red colours and enhanced passive fractions relative to the field. In addition, the formation mechanism of massive galaxies in clusters is redshift dependent: at z>2, massive galaxies must have assembled within 0.5 Gyr of them forming a significant fraction of their stars. So, in contrast to predictions from numerical simulations, few massive cluster galaxies formed via dry mergers, whereas at 2>z dry merging is a more important formation mechanism.