MacroKELLY lean: 42%
Will US CPI year-over-year inflation print below 3% at the next monthly release?
Horizon: 30 days
Private forecast journal
Make better predictions. Learn when reality proves you right or wrong.
Make predictions, compare with KELLY, and learn from how reality settles them. Start with a clear question, put a probability on it, compare with KELLY, and come back after the outcome settles to see your score.
Browse the guide without signing up. Your workspace stays private after you create an account.
Advanced - KELLY research mesh
Connected research labs stay read-only
Hypothesis, Science, and Memory Lineage labs feed KELLY calibration context. Your private journal never becomes their ground truth.
Live mesh status is exposed through /api/cognition-organ/pulse. Root page rendering does not wait on peer probes.
Explore forecasts
Browse starter questions before you sign up.
Landing preview uses the same explore feed ordering as the public browse page. Open Explore for filters and the full catalog.
Layout: Live proof first
PolicyKELLY lean: 38%
Will the Fed cut rates at least once before the end of this quarter?
Horizon: 90 days
TechnologyKELLY lean: 55%
Will a major AI lab ship a widely used open-weight model this month?
Horizon: 30 days
The forecast loop
From question to calibration score.
1
Ask
Pick a question you can score when reality arrives.
2
Estimate
Give your probability and compare it with KELLY.
3
Inspect
Read the evidence, base rate, and what would change your mind.
4
Commit
Keep your forecast private in your own journal.
5
Settle
Reality closes the outcome before scoring begins.
6
Learn
See your calibration score and what to improve next.
Where to start
Same KELLY, different jobs.
KELLY Superforecasting is your private journal. Ask KELLY handles broader conversations. Markets Lab handles market evidence. The proof shelf holds receipts.
What you get
Clear forecasts, visible proof, honest scores.
Your forecast stays separate from KELLY's forecast
You see the resolution rule before the outcome closes
Evidence, uncertainty, and base rates stay inspectable
Brier score and calibration update after settlement
Replays and learning stay tied to closed outcomes
You can ask and compare without writing KELLY's memory
Ask better questions
KELLY should explain its probability, not hide behind a number.
Ask inside KELLY Superforecasting when the question is about probability, calibration, and settlement. Use the full Ask KELLY app for broader personal conversations.
What evidence would make KELLY lower this forecast?
Which base rate is KELLY using, and how similar is it?
What is the settlement rule before reality arrives?
How did this outcome change calibration after closure?