What was missing was never the information.

Every founder discovers the same thing eventually. The decisions that reach you are the ones nobody below you could settle. The people you could ask are affected by the outcome. And the person who would be most useful, someone who would slow you down, ask what you were assuming, and open a possibility you had not considered, is usually not in your life at the moment you need them.

Why methods and not personalities

The obvious version of this product is five simulated famous executives. We built the other one, because the evidence points elsewhere. What has been shown to improve judgement, in the largest study run on the question, is training in specific procedures: state coherent probabilities, challenge the assumption, name the causal drivers, build the worst case. About an hour of it improved forecasting accuracy by roughly 10% and the gain persisted for years (Mellers, Tetlock et al., Good Judgment Project, sources).

Nothing in that literature suggests that borrowing someone's voice helps. So the board is built from methods, and which five you get depends on which ways of thinking you do not already use.

The best board is not the one that thinks like you. It is the one that covers what you skip.

What we are still proving

Two things, stated plainly because a page like this usually does not. We have not yet demonstrated that a session at month twelve beats a session on day one; that requires replaying held-out past decisions with and without the record, and we are building that test with the pilot cohort. And cognitive complementarity is our working hypothesis, defensible but not settled, as our method page sets out in more detail than is comfortable.

Five Peers is a service of Holco Invest, based in Paris. The decision record runs on HOLCO and is hosted in France.

Publication director: Pierre Coquard. Contact: contact@fivepeers.com