Independent Forecasts

When people make predictions without first discussing them together, their different perspectives and knowledge lead to more accurate combined forecasts than if they influenced each other first. If your team discusses market trends before sharing predictions, you'll likely end up with similar guesses that miss important insights—whereas separate forecasts capture diverse viewpoints and produce better collective wisdom. The key is sharing your thinking independently, then comparing results. Ask this: Do we share predictions before we share reasons?

Independent forecasts work because each person brings unique experiences, information, and mental models to the problem. When you predict separately, errors tend to cancel out rather than compound—someone's overoptimism balances another's caution, and one person's blind spot is another's area of expertise. The mathematical principle behind this is that averaging uncorrelated predictions reduces overall error more effectively than averaging similar ones.

In business settings, this means having team members write down sales projections, budget estimates, or risk assessments before any group discussion. A product team might have engineering predict technical feasibility, marketing assess customer demand, and finance evaluate costs—all independently. When these separate forecasts are combined, they often outperform what any single expert (or the group consensus) would produce.

In personal decisions, apply this by seeking multiple perspectives without first revealing your own thinking. If you're considering a job change, ask friends in different industries, your mentor, and family members for their independent assessments before sharing your leanings. You'll get more diverse—and more useful—input.

The danger is subtle: when people discuss before predicting, they anchor on each other's views, share convincing but incomplete information, and fall prey to social pressure toward conformity. This doesn't feel like bias—it feels like collaboration. But the result is a narrower range of predictions that collectively miss more than they catch. The solution is simple: forecast first, discuss second.

The mathematical foundation for independent forecasts comes from the diversity prediction theorem, formalized by Scott Page, which shows that collective error equals average individual error minus prediction diversity. When forecasts are independent, their errors are less correlated, increasing diversity and reducing collective error. Research by Larrick and Soll demonstrates that simply averaging multiple judgments typically improves accuracy, but this benefit erodes when forecasters influence each other before predicting.

• Page (2007): The diversity prediction theorem proves that collective predictive accuracy depends on both individual ability and prediction diversity, where mathematically, collective error equals average individual error minus diversity [1].
• Larrick and Soll (2009): Averaging multiple judgments consistently improves accuracy across a wide range of environments, yet people underutilize this strategy in favor of choosing between estimates [2].
• Lorenz et al. (2011): Social influence during forecasting reduces the wisdom of the crowd effect, as groups become more confident but less accurate when members can observe each other's predictions [3].
• Surowiecki (2004): The wisdom of crowds requires four conditions: diversity of opinion, independence, decentralization, and aggregation—without independence, crowds lose their predictive advantage [4].

Independence isn't always possible or desirable. Some problems require specialized knowledge that only a few people possess, making diverse forecasts impractical. Time constraints may prevent separate forecasting rounds. And certain decisions benefit from shared understanding and consensus-building over pure predictive accuracy. Additionally, completely independent forecasts can miss obvious information that everyone should reasonably know—sometimes a brief alignment on facts improves both individual and collective judgments. The key is distinguishing between sharing information versus sharing conclusions.