FirmFirm workspace
Experiment results
Blind forecast experiment

Forecast lab

Make and score independent event probabilities before comparison forecasts are revealed.

Your estimate stays independentYou see the exact question, scoring rule, deadline, and frozen evidence. Model, market, published MacroGuru, and other participants’ probabilities stay hidden until the experiment permits comparison.
Open questions
0accepting independent estimates
My forecasts
0latest saved versions
Resolved
0ready for scoring
Experiment status
Learningneeds more resolved outcomes
Owner or administrator

Add blind forecast question

Step 1 of 3 · QuestionMacroGuru generates custody identifiers, horizon math, and timestamps before the task is frozen.

Define the measurable event

One binary, independently resolvable claim.
Used for held-out calibration, not access control.
When the official outcome should be knowable.
Forecast window and related eventSet after the outcome deadlineSystem default

Forecasting opens now and normally stops before the outcome. Adjust only for a scheduled opening, earlier publication cutoff, or another claim about an existing event.

Defaults to now.
Defaults safely before resolution.
Choose an existing event only when this is another threshold or claim about the same underlying outcome.

Progress stays on this page until created, discarded, or signed out.
Blind human lane

Submit forecast

The single estimate used for proper scoring.
Uncertainty rangeOptional context; never used as a substitute pointOptional

Add a range only when it communicates material epistemic uncertainty. MacroGuru never imputes its midpoint as your scored probability.

Select only records that affected this estimate.

Record outcome

Defaults to now.
A reviewer must be able to reopen the exact result.