Find what could change your view. Research the unknowns in chat, compare evidence and saved model reasoning, and keep your team’s decisions connected to why they were made.
Begin with the investment judgment, then make the assumptions and the next research questions explicit.
An illustrative question for the next investment committee review. The team defines the investment horizon and what evidence would change its view.
Break the view into questions
Which inflation measure and threshold would change the thesis?
What would make the expected policy-rate path less credible?
Which range of ten-year yields would challenge the view?
An investment view with defined supporting questions, a review horizon and evidence that could change it.
A Yes / No condition with explicit resolution criteria.
Numeric questions and ranges for rates, inflation, valuation or another driver, with units, horizons and saved estimates.
A numeric estimate with units and a defined period.
Compare several plausible pathways and the conditions behind each for questions that need more than a binary outcome.
Several plausible paths and the conditions behind each.
A view moves through research, debate and decisions. Keep the people, evidence and rationale connected so the next review can start from what actually happened.
Start with a question in chat. Crene helps connect the investment view to supporting questions, dated research, saved model reasoning and review decisions. The structured record makes the assumptions and subsequent changes easier to inspect.
They show available models’ saved assessments of a defined question, with their reasoning and dates. Differences help the team examine competing assumptions. Disagreement is not a confidence interval or proof of forecast accuracy.
Workspace members can assign an owner and reviewer, set a task due date, comment and mention colleagues in the app. A review task, a team decision, a personal forecast and an event outcome are separate records.
Requests depend on workspace access, selected models and usage limits. On-demand assessments show a cost preview and require confirmation. Recurring monitoring has its own controls; opening saved results does not run another assessment.
Resolved binary questions can be evaluated using Brier scores and calibration analysis. These evaluate forecasts against outcomes; they do not by themselves demonstrate trading returns. The Data and Methodology pages describe the measurement record and its limitations.
Private evaluation workspaces are offered to selected individuals by invitation. Do not enter confidential research into a shared sample workspace.
Use a six-week private review to define one thesis, examine the evidence and model disagreement, and retain the team’s review decisions.
One thesis. One team. Six weeks.
Workspace access is arranged by invitation after the scope review. Sending an inquiry does not create an account. Share a non-confidential summary to start.