On the later evolution of observationally selected protocluster candidates at $z\gtrsim 5$

Recent observations have revealed numerous protocluster candidates at $z\gtrsim 5$, yet whether these systems will eventually evolve into today’s galaxy clusters remains an open question. Using the FLAMINGO simulations – resolving protocluster cores up to $z\simeq 10$ – we track the later evolution of observationally selected protocluster candidates, comparing three selection methods against observational samples. The observed number density falls between our mass-selected and abundance-matched samples, implying that current searches pick up both genuine cluster progenitors and significant interlopers that will not reach cluster masses by $z=0$. We find that candidates at $z\gtrsim 5$ are heavily clustered, hosting 2$-$10 neighbors within 10,cMpc. Consequently, a candidate with a neighbor at 5,cMpc (10,cMpc) faces a $\gtrsim 50\mathrm{\%}$ ($\gtrsim 30\mathrm{\%}$) probability of later merging into a larger system, mostly at $z\lesssim 2$. The merger count converges beyond $\sim 10$,cMpc, pointing to a fundamental scale in structure formation. Observations show markedly weaker clustering than our simulations predict, suggesting clustering offers a currently overlooked diagnostic. Each candidate undergoes roughly 2$-$6 later major mergers, mostly with systems too small to be recognized as massive at the selection epoch. Hence, relying solely on high-$z$ mass and galaxy overdensity to forecast a candidate’s fate is prone to severe scatter and systematic error. A robust identification of true cluster progenitors demands a total mass sum of galaxies down to the faintest levels within a 10,cMpc radius. Upcoming surveys with both depth and area will be key to reliably linking high-$z$ protocluster candidates to their ultimate destiny.