The Science of Board Decisions: Theory, Math, and Practice

Not legal advice. This page collects what decision theory and group-decision research say about how a board can decide well, with every finding linked to its source (full list at the end). Where we give Prepared Board's own view, it is labeled Prepared's view (opinion). Product claims stay within Facts; Prepared Board does not compute expected values, probabilities, or any decision or outcome score.

A board is a small group making infrequent, high-stakes, uncertain decisions, usually on information prepared by the people it oversees. That combination is exactly where the research says individual judgment is weakest and group process matters most. The good news: the fixes are mostly cheap.

Part I — Decision theory for directors

1. Expected value, and why boards are not risk-neutral

The expected value of an option is the probability-weighted average of its outcomes:

EV(option) = Σ  p(outcome i) × value(outcome i)

Worked example. A board is asked to approve a $2.0M product launch. Management estimates a 30% chance it returns $10M, a 50% chance it returns $2M, and a 20% chance it returns nothing.

EV(launch) = 0.30 × 10.0 + 0.50 × 2.0 + 0.20 × 0.0 − 2.0
           = 3.0 + 1.0 + 0.0 − 2.0
           = +2.0  ($M)
EV(do nothing) = 0

Expected value favors launching. But von Neumann and Morgenstern's Theory of Games and Economic Behavior (1944) showed that rational choice under uncertainty maximizes expected utility, not expected dollars — and utility is where risk preferences live. A board that would be wiped out by the 20% branch is entitled to weigh it more heavily than the arithmetic does. The useful board question is not "what is the EV?" but "what are the branches, how likely is each, and which can we not survive?"

2. Decision trees and the value of information

Howard Raiffa's Decision Analysis (1968) popularized drawing a decision as a tree — decisions, chance events, payoffs — and "rolling back" from the end to find the best first move. Its most practical lesson for boards is that information has a price you can estimate.

Worked example. Launch now, or run a $0.3M pilot first? There is a 40% chance demand is strong (launch nets +$6M) and a 60% chance it is weak (launch nets −$2M). Suppose the pilot reliably reveals which.

Launch now:      0.4 × 6 + 0.6 × (−2)            = 2.4 − 1.2 = +1.2
Pilot, then launch only if strong:
                 0.4 × 6 + 0.6 × 0 − 0.3         = 2.4 − 0.3 = +2.1
Value of perfect information = 2.4 − 1.2        = 1.2

The pilot is worth buying at any cost below $1.2M (less if the pilot is imperfect). The same logic tells a board when to defer a decision and when deferral is just avoidance: deferral is worth it only if the extra information could change the choice.

3. Base rates and the outside view

Kahneman & Tversky (1973) showed that people making predictions tend to neglect base rates — how often things like this happen in general — in favor of how well a specific case "fits" a story. Kahneman & Lovallo (1993) applied this to organizations: planners take an inside view (this project, its plan, its obstacles) and produce bold forecasts; the corrective is an outside view that asks how a reference class of similar projects actually turned out.

For a board, the outside view is a single question asked before discussing the plan: "Of the last N decisions like this — ours or anyone's we can find — how many hit their targets, and by how much did they miss?" Prepared Board does not supply reference-class data; it can show a board its own history (see Part IV).

4. Bayesian updating: how much should one good quarter move us?

Bayes' rule, from Thomas Bayes's posthumously published essay (1763), tells you how far evidence should move a belief:

P(H | E) =            P(E | H) × P(H)
           ─────────────────────────────────────────
           P(E | H) × P(H)  +  P(E | not H) × P(not H)

Odds form:  posterior odds = prior odds × likelihood ratio
            likelihood ratio = P(E | H) / P(E | not H)

Worked example. The board thinks there is a 30% chance a new market's demand is durable (H). A strong quarter (E) would happen 80% of the time if demand is durable, but 40% of the time anyway (a one-off customer, pull-forward).

P(H | E) = (0.8 × 0.3) / (0.8 × 0.3 + 0.4 × 0.7)
         = 0.24 / 0.52
         ≈ 0.46

Odds form: (0.3 / 0.7) × (0.8 / 0.4) = 0.43 × 2 ≈ 0.86  →  0.86 / 1.86 ≈ 0.46

One strong quarter should move the board from 30% to about 46% — meaningful, but nowhere near "proven." Two habits follow: write down the prior before the data arrives, and ask "how likely is this evidence if we're wrong?" (the denominator), which is the question presentations rarely answer.

5. Anchoring

In Tversky & Kahneman's (1974) classic demonstration, participants spun a wheel rigged to stop at 10 or 65, then estimated the percentage of African countries in the United Nations. Median estimates were 25 for those who saw 10 and 45 for those who saw 65 — an obviously irrelevant number pulled judgments toward itself. In a boardroom, the first number on the slide (management's valuation, last year's budget, the banker's range) is an anchor. Asking directors for their own estimate before the number is shown is the simplest defense.

6. Calibration: keeping score on judgment

The Brier score, introduced for weather forecasts by Brier (1950), measures how good probabilistic forecasts are:

Brier score = (1/N) × Σ (forecast probability − outcome)²     outcome = 1 if it happened, 0 if not
              0 is perfect; always saying 50% scores 0.25

Example: "70% likely" and it happened   → (0.7 − 1)² = 0.09
         "90% likely" and it didn't     → (0.9 − 0)² = 0.81

In a large geopolitical forecasting tournament, Mellers et al. (2014) found that training in probabilistic reasoning, working in teams, and tracking top performers into elite teams all improved forecast accuracy. Boards almost never score their own forecasts. Recording an explicit expectation at decision time ("we expect X by date Y") is the precondition for ever doing so.

Part II — The group science of boards

7. Condorcet's jury theorem — and the assumptions that break in boardrooms

The Marquis de Condorcet's Essai sur l'application de l'analyse à la probabilité des décisions rendues à la pluralité des voix (1785) gives the mathematical case for deciding by majority. If each of n voters (odd) independently picks the correct option of two with probability p:

P(majority correct) = Σ_{k = (n+1)/2}^{n}  C(n, k) × p^k × (1 − p)^(n − k)

Directors (n)    p = 0.60    p = 0.55    p = 0.45
      1           0.600       0.550       0.450
      3           0.648       0.575       0.425
      7           0.710       0.608       0.392
      9           0.733       0.621       0.379
     11           0.753       0.633       0.367

When each director is a bit better than a coin flip, a larger majority is more reliable than any one of them. But the same formula runs in reverse: if directors are on average worse than a coin flip on a question (p = 0.45), adding votes makes the majority more reliably wrong.

The theorem rests on assumptions that boards routinely violate:

• Independence. If nine directors all formed their view from the same management deck and the chair's framing, they are closer to one signal than nine. Ladha (1992) showed majorities can still beat individuals when votes are correlated, but only if the average correlation is low enough — a much stricter condition in small groups.
• Sincere voting. Austen-Smith & Banks (1996) showed that even when everyone shares the goal, voting one's private information sincerely need not be rational: a voter may do better by conditioning on what others' votes imply.
• Competence and a shared objective. Grofman, Owen & Feld (1983) extend the theorem to voters of differing competence; the majority advantage depends on average competence above one-half on that question.
• A binary, well-posed question. Many board questions are multi-option or ill-defined until the brief is written.

Practical reading: a board gets the Condorcet benefit only if it protects independent judgment before discussion and asks questions the directors are actually competent to judge.

8. Wisdom of crowds vs. groupthink

At a 1906 livestock fair, Francis Galton collected 787 usable guesses of an ox's dressed weight; the median guess was 1,207 lb against an actual 1,198 lb (Galton 1907, Nature). James Surowiecki's The Wisdom of Crowds (2004) argues that crowds are wise under four conditions: diversity of opinion, independence, decentralization, and a way to aggregate.

Irving Janis's Victims of Groupthink (1972) — expanded in 1982 as Groupthink: Psychological Studies of Policy Decisions and Fiascoes — describes the opposite: cohesive groups that suppress dissent, using cases such as the Bay of Pigs. Janis listed symptoms including an illusion of invulnerability, collective rationalization, pressure on dissenters, self-censorship, an illusion of unanimity, and self-appointed "mindguards." The honest caveat: Esser (1998), reviewing 25 years of research, found the empirical support mixed — laboratory tests often fail to confirm the full model, while case studies offer more support. Treat groupthink as a useful lens, not a law.

The underlying conformity effect is better established. In Asch's (1956) line-judgment experiments, participants facing a unanimous but plainly wrong majority went along with it on roughly a third of critical trials, though about a quarter never conformed at all.

9. The hidden-profile problem

Stasser & Titus (1985) gave each member of a group part of the information about candidates. Pooled, the information pointed to the best choice; but discussion dwelled on what members already shared, and groups picked the initially favored, inferior option. A meta-analysis of 65 studies by Lu, Yuan & McLeod (2012) found groups mentioned far more common than unique information and that groups facing a hidden profile were about eight times less likely to find the best solution than groups given full information.

Boards are hidden-profile machines: the audit chair knows about the control weakness, the newest director heard something from a customer, the CEO knows the pipeline's soft spots. None of it surfaces unless the process asks for it before the room converges.

10. Information cascades and speaking order

Bikhchandani, Hirshleifer & Welch (1992) and Banerjee (1992) showed how sequential decisions can produce cascades: once a few people have chosen the same way, a later person can rationally ignore their own private information and follow, so the group stops learning — and the cascade can be fragile, flipping on a small piece of public news. In a boardroom, the sequence is the speaking order. If the chair and the lead investor speak first, the remaining directors' private signals may never be heard. Reversing the order (newest or most junior first), or collecting views in writing before anyone speaks, preserves information.

Part III — Methods that follow from the research

A one-page board protocol (Prepared's view, opinion)

Prepared's view (opinion): none of the above needs software. A board that does these seven things will, we believe, decide better than one with a beautiful portal that doesn't:

1. Write the question and the options before the meeting — including "do nothing" and "defer, and what would we learn."
2. State the base rate for this kind of decision, even roughly, and where it came from.
3. Collect each director's independent view in writing (lean, key concern, estimate) before the first number is presented.
4. Run a ten-minute pre-mortem: "It's two years from now and this failed. Why?"
5. Reverse the speaking order — newest or most junior directors first; the chair last.
6. Record conflicts and recusals before discussion, not after.
7. Write down what you expect, by when, and who will report — then come back and review the decision process separately from the result.

Part IV — What Prepared Board supports today (and what it doesn't)

Only live routes are linked; signed-in links work in the seeded demo boards. Limits are stated plainly.

Where to go next

• History of how boards got here: A short history of boards.
• What the evidence does and doesn't show about boards and outcomes: Do boards matter?.
• Related essays: The decision record is the product, Outcome monitors after the gavel.

Try it in a demo board

Open Northlight Robotics, a seeded company demo — not a real issuer — at Try a board, or sign in as a sample board (see /sample-decision) (demo password on Facts). Open a decision to see the Decision Brief, ballots, recusals, and monitors; then open /app/decisions/outcomes. What is not live (SSO, billing, computed scores, and more) is listed on Facts.

Sources

1. von Neumann, J., & Morgenstern, O. (1944). Theory of Games and Economic Behavior. Princeton University Press. press.princeton.edu
2. Raiffa, H. (1968). Decision Analysis: Introductory Lectures on Choices Under Uncertainty. Addison-Wesley. Google Books
3. Kahneman, D., & Tversky, A. (1973). "On the Psychology of Prediction." Psychological Review 80(4): 237–251. doi:10.1037/h0034747
4. Kahneman, D., & Lovallo, D. (1993). "Timid Choices and Bold Forecasts: A Cognitive Perspective on Risk Taking." Management Science 39(1): 17–31. doi:10.1287/mnsc.39.1.17
5. Bayes, T. (1763). "An Essay towards Solving a Problem in the Doctrine of Chances." Philosophical Transactions of the Royal Society of London 53: 370–418. doi:10.1098/rstl.1763.0053
6. Tversky, A., & Kahneman, D. (1974). "Judgment under Uncertainty: Heuristics and Biases." Science 185(4157): 1124–1131. doi:10.1126/science.185.4157.1124
7. Brier, G. W. (1950). "Verification of Forecasts Expressed in Terms of Probability." Monthly Weather Review 78(1): 1–3. doi:10.1175/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2
8. Mellers, B., Ungar, L., Baron, J., Ramos, J., Gurcay, B., et al. (2014). "Psychological Strategies for Winning a Geopolitical Forecasting Tournament." Psychological Science 25(5): 1106–1115. doi:10.1177/0956797614524255
9. Condorcet, M. J. A. N. de Caritat, Marquis de (1785). Essai sur l'application de l'analyse à la probabilité des décisions rendues à la pluralité des voix. Paris: Imprimerie royale. archive.org
10. Ladha, K. K. (1992). "The Condorcet Jury Theorem, Free Speech, and Correlated Votes." American Journal of Political Science 36(3): 617–634. doi:10.2307/2111584
11. Austen-Smith, D., & Banks, J. S. (1996). "Information Aggregation, Rationality, and the Condorcet Jury Theorem." American Political Science Review 90(1): 34–45. doi:10.2307/2082796
12. Grofman, B., Owen, G., & Feld, S. L. (1983). "Thirteen Theorems in Search of the Truth." Theory and Decision 15(3): 261–278. doi:10.1007/BF00125672
13. Galton, F. (1907). "Vox Populi." Nature 75: 450–451. doi:10.1038/075450a0
14. Surowiecki, J. (2004). The Wisdom of Crowds. New York: Doubleday. archive.org
15. Janis, I. L. (1972). Victims of Groupthink: A Psychological Study of Foreign-Policy Decisions and Fiascoes. Boston: Houghton Mifflin. 2nd ed. (1982) Groupthink: Psychological Studies of Policy Decisions and Fiascoes. archive.org
16. Esser, J. K. (1998). "Alive and Well after 25 Years: A Review of Groupthink Research." Organizational Behavior and Human Decision Processes 73(2–3): 116–141. doi:10.1006/obhd.1998.2758
17. Asch, S. E. (1956). "Studies of Independence and Conformity: I. A Minority of One against a Unanimous Majority." Psychological Monographs: General and Applied 70(9): 1–70. doi:10.1037/h0093718
18. Stasser, G., & Titus, W. (1985). "Pooling of Unshared Information in Group Decision Making: Biased Information Sampling during Discussion." Journal of Personality and Social Psychology 48(6): 1467–1478. doi:10.1037/0022-3514.48.6.1467
19. Lu, L., Yuan, Y. C., & McLeod, P. L. (2012). "Twenty-Five Years of Hidden Profiles in Group Decision Making: A Meta-Analysis." Personality and Social Psychology Review 16(1): 54–75. doi:10.1177/1088868311417243
20. Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). "A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades." Journal of Political Economy 100(5): 992–1026. doi:10.1086/261849
21. Banerjee, A. V. (1992). "A Simple Model of Herd Behavior." Quarterly Journal of Economics 107(3): 797–817. doi:10.2307/2118364
22. Klein, G. (2007). "Performing a Project Premortem." Harvard Business Review, September 2007. hbr.org
23. Mitchell, D. J., Russo, J. E., & Pennington, N. (1989). "Back to the Future: Temporal Perspective in the Explanation of Events." Journal of Behavioral Decision Making 2(1): 25–38. doi:10.1002/bdm.3960020103
24. Schweiger, D. M., Sandberg, W. R., & Ragan, J. W. (1986). "Group Approaches for Improving Strategic Decision Making: A Comparative Analysis of Dialectical Inquiry, Devil's Advocacy, and Consensus." Academy of Management Journal 29(1): 51–71. doi:10.2307/255859
25. Sunstein, C. R., & Hastie, R. (2015). Wiser: Getting Beyond Groupthink to Make Groups Smarter. Boston: Harvard Business Review Press. store.hbr.org
26. Kahneman, D., Sibony, O., & Sunstein, C. R. (2021). Noise: A Flaw in Human Judgment. New York: Little, Brown Spark. archive.org
27. Kahneman, D., Lovallo, D., & Sibony, O. (2019). "A Structured Approach to Strategic Decisions." MIT Sloan Management Review, Spring 2019. sloanreview.mit.edu
28. Kahneman, D., Lovallo, D., & Sibony, O. (2011). "Before You Make That Big Decision…" Harvard Business Review, June 2011. hbr.org
29. Baron, J., & Hershey, J. C. (1988). "Outcome Bias in Decision Evaluation." Journal of Personality and Social Psychology 54(4): 569–579. doi:10.1037/0022-3514.54.4.569
30. Fischhoff, B. (1975). "Hindsight ≠ Foresight: The Effect of Outcome Knowledge on Judgment under Uncertainty." Journal of Experimental Psychology: Human Perception and Performance 1(3): 288–299. doi:10.1037/0096-1523.1.3.288
31. Lovallo, D., & Sibony, O. (2010). "The Case for Behavioral Strategy." McKinsey Quarterly, March 2010. mckinsey.com
32. Internal Revenue Service (2024). Instructions for Form 1023 (Rev. December 2024), Appendix A: Sample Conflict of Interest Policy. irs.gov PDF
33. State of Delaware (2026). Title 8, Delaware Code, §144 ("Interested directors; quorum"), text as published on the Delaware Code Online, accessed 5 October 2026. Check current text with counsel; §144 has been amended over time. delcode.delaware.gov
34. Smith v. Van Gorkom, 488 A.2d 858 (Del. 1985). Justia