The real world is fraught with risk.
Forecasting with empirical data - the #NoEstimates approach to estimating - ignores this fact. In the NE version of forecasting there is no place for uncertainty, which is why No Estimates advocates assert Forecasting is not estimating. In this version of Estimating without calling it Estimating, the past behaviour represents the future behaviour, with not adjustments for reducible and irreducible uncertainties that create risk to the project's success.
When we try to plan, knowing we cannot predict the future precisely, the mismatch between the planning and the real world, creates confusion. This confusion creates biases and distortions, resulting in padded estimates and undue optimism as well as misuse and even dysfunction of the management in the presence of uncertainty.
There needs to be an approach to planning and forecasting that is connected with reality, an approach that acknowledges uncertainty from the beginning and continues to manage in the presence of that uncertanty throughout the life of the project.
This approach mandates estimating all our work that not de minimis. This work is driven by uncertainty - reducible and irreducible.
This uncertainty creates risk and Risk Management is How Adults Manage Projects - Tim Lister.
Practical Risk Assessment for Project Management, Stephen Grey, Wiley Series in Software Engineering Practice, 1995. Nothing has changed since these words. We still live, operate, and manage in the presence of uncertainty. Here's how we manage in our software intensive system of systems domain. Your domain may be different, but the principles of managing in the presence of uncertainty are the same.
Here's the now familiar briefing for how we management in the presence of uncertainty in our Software Intensive System of Systems domain.