A scientific framework quantifies the reliability of individual forecasts in uncertain environments and identifies mechanisms for improving predictive accuracy. Philip Tetlock is a key figure who systematically studied expert judgment, demonstrating that most experts perform only marginally better than chance in long-term predictions within their own domains. His work established methods for evaluating forecasting skill by breaking down predictions into specific, verifiable propositions and assigning probabilities to outcomes.
This field defines the conditions under which human prediction can be disciplined and how biases commonly distort assessment. The boundaries of this framework apply to explicit, testable predictions about future events and do not extend to general strategic visions or qualitative assessments without specific, measurable outcomes.
This dimension addresses how the quality of individual predictions can be objectively measured beyond simply whether an event occurred. Forecasts are assessed not only on their accuracy but also on their calibration, meaning the extent to which stated probabilities correspond with the actual frequency of events. A forecaster predicting an event with 70% confidence should see that event occur about 70% of the time across many predictions.
Resolution, another component, indicates how well a forecaster differentiates between more and less likely outcomes, assigning high probabilities to events that happen and low probabilities to those that do not. When a founder operates without these distinctions, for example by mistaking high confidence for high accuracy, the cost in misallocated capacity becomes substantial. This can lead to wasted hours pursuing initiatives with an inflated perceived chance of success, or neglecting preparation for contingencies that were genuinely probable but dismissed by overconfident judgment.
The mental effort and inherent biases involved in prediction impose a measurable burden on an individual. Philip Tetlock’s research highlights how common cognitive biases, such as confirmation bias, overconfidence, and anchoring, systematically distort judgments even among highly informed people. For a founder carrying the work of an operator, this means expending substantial attention capacity to counteract these natural tendencies, or simply enduring the energy cost of making decisions under the influence of skewed forecasts.
The ‘fox’ and ‘hedgehog’ distinction identifies different cognitive styles, where foxes use multiple frameworks and adapt their thinking to new evidence, while hedgehogs stick to one grand theory. A founder leaning toward the hedgehog style, despite potential for deep domain expertise, risks higher cognitive load when reality challenges their core narrative, incurring costs in adaptive capacity and potentially delayed course corrections.
Adopting a quantitative approach to forecasting requires a deliberate shift from vague qualitative statements to specific probabilities. Instead of declaring an initiative “likely” to succeed, a founder engaging with this method assigns a precise percentage, such as a 65% chance of success within a defined timeframe. This forces greater clarity about underlying assumptions and allows predictions to be systematically tracked and evaluated using metrics like the Brier score, which quantifies the difference between stated probabilities and actual outcomes, penalizing both inaccuracy and poor calibration.
The discipline of stating explicit probabilities enables the aggregation of diverse individual forecasts, often combining several independent estimates to produce a collective prediction that often surpasses any single predictor’s ability, particularly when individual biases average out. This method reduces the reliance on intuition alone and introduces a mechanism for continuously refining a founder’s understanding of risk and opportunity, moving beyond a simple binary ‘yes’ or ‘no’ outcome. The cost involves focused attention and consistent training to apply this thinking, which demands a shift in mental models for assessing uncertainty, but it yields greater robustness in planning and decision-making by making assumptions transparent and testable.
One word went out and arrived as clarity with one person and as an accusation with another. What a word means on arrival belongs to the listener's default, and a counting word said where nobody counts moves the relation outward,
The terms are written where they belong. Each field holds three desks, and every entry a desk writes raises the terms it uses, each term given a meaning, a mechanism and the places it appears. The record on these pages stays first person and hand made; the fields grow the nomenclature.
