About
Pathspace Lab is an independent research studio exploring risk, uncertainty, and complex systems through simulation.
Most analysis is conditioned on a single realized history: the one sequence of events that actually happened. Pathspace Lab starts from a different question — not what happened, but how much of what we believe depends on the particular history we happened to observe? Rather than forecasting forward or backtesting backward, the work here looks sideways: generating ensembles of plausible alternative histories and asking whether conclusions — portfolio allocations, risk estimates, factor exposures — hold up across them, or whether they’re artifacts of one lucky (or unlucky) sequencing of events.
Where the idea came from
The name comes from mathematics and physics, where a pathspace is the collection of all possible paths a system can take through time. My interest in the idea predates finance. In graduate school I studied stochastic processes in biophysical systems, where individual molecules, proteins, and ion channels evolve along many possible trajectories rather than a single predetermined one — understood through ensembles of trajectories, not one realized path.
Later, building equity risk models, I kept running into the same shape of question wearing different clothes: markets, too, produce one observed history, but countless others were equally plausible given the same underlying dynamics. That parallel — ensembles in biophysics, single realized histories in finance — is what Pathspace Lab grew out of.
Approach
Work here favors conservative, information-preserving perturbations over speculative scenario design: resampling and reshuffling what’s already in the data (block bootstraps of returns, alternative time-orderings) rather than injecting new assumptions about how markets behave. The aim isn’t predictive accuracy — it’s understanding how fragile or robust a given conclusion is to the history that happened to unfold.
For the full story of how this project started, read the founding essay:
Get in touch
If these questions interest you, I’d welcome the opportunity to connect, collaborate, or simply exchange ideas.