About Crene
Built to make investment judgment inspectable.
Crene turns investment theses into living decision records. Research what could change the view, examine the assumptions behind model disagreement, and keep the team’s review decisions connected to their evidence.
- 01The question and its assumptions
Definitions, horizons and the conditions that matter.
- 02The evidence and model reasoning
Dated sources, available estimates and disagreement.
- 03The team’s decision and rationale
Maintain, revise or retire the view, with a reason.
- 04The prior states and history
What the team believed and how that view changed.
Forecast measurement is disclosed separately from the structure of a decision record.
01 · The mission
Know what your view depends on.
An investment review should expose what remains uncertain and what deserves attention next. Crene connects that work to the question, the evidence and the decisions your team records.
- 01
Research
Start in chat. Investigate a question, inspect dated sources and saved model reasoning, and identify assumptions or missing evidence that deserve a closer look.
- 02
Review
Use binary questions, numeric drivers or scenarios to frame the uncertainty. Review definitions, sources and horizons, compare available model views, and assign the follow-up.
- 03
Decide
PMs, risk teams, CIOs and investment committees can record whether to maintain, revise or retire a view and why. Personal forecasts and recorded event outcomes stay separate from that team decision.
02 · Our principles
What Crene believes.
The view, the evidence, the model estimate and the human decision each need a clear place in the record.
01
A thesis should have a decision record.
The question, its definition changes, the evidence considered and the team’s decision history should remain connected.
02
Disagreement should be visible.
Read the reasoning and uncertainty behind available model estimates. Use competing assumptions and missing evidence to shape the next research question.
03
Calibration should be disclosed.
Eligible resolved binary forecasts can be scored against recorded outcomes. The sample, exclusions and date of the record matter as much as the headline score.
04
Structure and scoring should stay separate.
A scenario pathway, a numeric driver, a personal probability and a team decision describe different things. Their evaluation should respect those differences.
03 · Founder
Stephen Lee
Founder & CEO
Stephen Lee worked in institutional finance at Goldman Sachs and Crédit Agricole before moving into software engineering.
He built Crene to help investment teams examine what their view depends on, notice what deserves further research, and retain the evidence and reasoning behind their decisions.
stephen@crene.com ↗04 · Accountability
Keep the record inspectable.
The historical public corpus and private workspace forecasts have different scopes. Crene makes those distinctions visible so the measurement record can be evaluated on its own terms.
Results have a scope.
Public historical scores do not establish the performance of a private workspace. Private scoring needs eligible dated forecasts and recorded outcomes with evidence.
Inspect the measurement record →Evidence requires review.
Source links and saved model reasoning give the team material to inspect. Retrieval alone does not establish that a claim is correct or that a source supports it.
Read the methodology →People remain accountable.
Owners and reviewers coordinate the work. On-demand assessments require their own confirmation, monitoring has separate controls, and human decisions remain explicit.
Explore a private review →Start with a real investment view, the decision it affects, and the evidence your team already uses.
Review a thesis →