Software estimation is fundamentally flawed because developers instinctively estimate the shortest possible time rather than the realistic time to do things right. This creates a blame culture where padding estimates feels taboo, time pressure forces shortcuts, and unmanaged technical debt accumulates. Ternary estimates (optimistic/expected/pessimistic) rarely work in practice. The real cost of estimation is that time pressure stops developers from solving problems properly — they reach for quick fixes instead of evaluating the full solution space. This compounds into legacy systems that are hard to change, ultimately making the business itself inflexible. A better approach frames estimates as a function of time, complexity, and risk, making all three dimensions explicit rather than collapsing everything into a single number.
