\subsection{Problem: Exogenous $P$ / Exogenous Informativeness}\label{sec:intro-problem}
% Prior work treats success probability and informativeness as given; no lever for founders to design learning capacity.
This precision parameter exposes a blind spot in entrepreneurship theory. Prior research
treats either \textbf{venture success probability} $P$ (probability of both selling and
delivering on venture's promise) or \textbf{informativeness} (how responsive beliefs are to
market information) as exogenous, yet the Tesla-Better Place divergence shows both are
endogenous to promise precision $\tau$. If $P$ is fixed, all ventures share identical odds, reducing survival differences to luck or founder talent. If informativeness is fixed, all ventures learn at identical rates from informational inflow, leaving heterogeneous adaptation unexplained.
Without a mechanism to preserve learning loops—when market signals update beliefs versus
when they don't—theory cannot explain why some ventures pivot while others march blindly.
Meanwhile, founders miss the most strategic variable of all: precision $\tau$, which
governs whether they can learn at all \citep{steen2017formal}.