To address a crucial point immediately: I agree with the professor that the future is not merely uncertain but inherently unknowable. However, some situations are more unknowable than others, and prediction need not achieve perfect accuracy. Being somewhat less wrong can still provide value. Even so, the potential pitfalls of prediction are numerous and serious.

These problems fall into three distinct categories: limitations in the forecasting methods and models themselves, human behavioral biases interfering with the prediction process, and communication breakdowns between forecasters and decision-makers.

Prediction methods often rely on past data and historical patterns, assuming previous relationships will persist—an assumption that frequently fails during periods of rapid change or unprecedented circumstances. Forecasting exercises typically concentrate on specific variables while neglecting broader contexts, blinding forecasters to crucial interconnected factors. When relatively unlikely events have disproportional consequences, prediction failure is huge. The Great Financial Crisis of 2008, chronicled by Nassim Nicolas Taleb in “The Black Swan,” exemplifies this dynamic. Most forecasting efforts focus on short time horizons, failing to account for structural changes or rare but high-impact events. Even well-built models often rest on incorrect assumptions about causal relationships, while reality’s complexity includes non-linear effects and emerging properties that defy prediction.

A second problem other than methods and models are the people running them. Human judgment suffers from predictable cognitive limitations: confirmation bias, overconfidence, groupthink, reliance on mental shortcuts, and emotional interference with rational analysis. Expert overconfidence proves particularly problematic, as domain knowledge can increase certainty without improving accuracy. Many forecasters employ deliberately vague language that makes post-hoc accuracy assessment impossible, allowing poor forecasters to claim vindication regardless of outcomes. When forecasting becomes institutionalized, perverse incentives emerge—forecasters may cherry-pick data or adjust methodologies to produce desired results, while organizations prefer conventional failures over original mistakes.

Finally as predictions get shared with leaders to inform some decision, they typically receive them without understanding underlying reasoning, methodology, or confidence intervals, leading to misplaced confidence in uncertain projections. Few forecasters excel at communicating inherent uncertainty, while decision-makers rarely conduct sensitivity analyses, instead treating point estimates as gospel. Qualities valued in organizational leaders—confidence, decisiveness, clear vision—directly conflict with intellectual habits that produce accurate forecasts: probabilistic thinking, intellectual humility, and comfort with admitting error. Organizational pressures favor simple, definitive answers over nuanced, probabilistic assessments. Once a decision is taken, it can lead an organization to ignore other possible outcomes and let a false sense of security prevent preparation for contingencies. And that is a point where there’s a lot to be said in favor of scenario planning.