The Base Rate Blindspot: Why You Keep Ignoring the Most Useful Number in the Room

There's a number that exists for almost every decision you'll make. It sits quietly in actuarial tables, historical records, and published research. It predicts outcomes with reasonable accuracy before you've gathered a single piece of case-specific evidence. And most people ignore it almost completely.

Photo by Waskyria Miranda on Pexels.

That number is the base rate: the frequency with which something happens across a reference population, before you layer in anything specific about your situation.

Consider a straightforward example. About 90% of restaurant startups fail within their first year. That's a base rate. When an entrepreneur pitches their new concept, full of genuine differentiators and a compelling origin story, the investors listening will typically evaluate the pitch on its own merits. They'll think about the team, the market timing, the unit economics. What almost no one does is start by anchoring hard to that 90% figure and working forward from there. The specific drowns out the statistical.

Psychologists call this the representativeness heuristic. We judge probability by how well something matches our mental prototype of success (or failure), not by how often outcomes actually occur in similar situations. A confident founder with a well-rehearsed deck matches our prototype of a winner. The base rate doesn't match anything; it's just a number on a page.

This matters enormously in high-stakes decisions.

Daniel Kahneman documents a vivid personal example in Thinking, Fast and Slow. He was advising a curriculum development team that had been working together for about a year. He asked each member to independently estimate how long the project would take to complete. Answers clustered around 18 to 24 months. Then he asked the most experienced member: of similar projects you've seen, what fraction finished on time, and how long did they typically take? The answer was sobering. Most comparable projects took 7 to 10 years. None had finished in under 7. The team's inside view had completely eclipsed a perfectly accessible outside view. They were not ignoring the base rate out of arrogance; they genuinely hadn't thought to ask for it.

The fix sounds obvious in retrospect: ask for the outside view first, then adjust for your specific situation. In practice, this requires reversing the natural order of attention.

Here's a rough process that helps:

graph TD
    A[Define the decision or forecast] --> B{Find a reference class}
    B --> C[Look up the base rate for that class]
    C --> D[Anchor your estimate to the base rate]
    D --> E[Identify specific factors that differentiate your case]
    E --> F[Adjust from the anchor, with discipline]
    F --> G[Final probability estimate]

The hardest step is B. Choosing the right reference class is genuinely difficult. Too broad, and the base rate loses relevance. Too narrow, and you don't have enough historical cases to produce a reliable frequency. The reference class problem (which we've covered separately) lives right here. But even an imperfect reference class beats having no anchor at all.

Where does this show up most destructively in organizational life? Three places keep recurring.

Mergers and acquisitions. Studies consistently show that roughly 70-80% of M&A deals fail to create value for the acquiring firm's shareholders. Individual deal teams almost universally believe their transaction will be in the successful minority. The synergy projections, the strategic rationale, the cultural fit assessments: all of it crowds out the baseline expectation that the deal will disappoint.

IT project overruns. The Standish Group's CHAOS Report has tracked software project outcomes for decades. Large projects fail to deliver on time and on budget at rates that have remained stubbornly high across different methodologies, industries, and eras. Each individual project team believes their planning is more rigorous than the average case. Some are right. Most aren't.

Medical prognosis. Physicians are systematically overoptimistic about patient survival when they have a personal relationship with the patient. The base rate for a given diagnosis predicts outcomes better than the clinician's intuitive estimate in a substantial fraction of cases. This isn't a knock on medical expertise; it's a property of how human judgment operates under conditions of emotional investment.

The useful reframe is this: base rates tell you what the world does on average when it encounters situations like yours. Your job is to find genuine, specific reasons to deviate from that average, and then to deviate by proportionate amounts. Not to ignore the average because your situation feels special.

Almost every situation feels special from the inside. That's exactly the problem.