Project Data
Forecast questions are developed in SciCast Spark and published to and forecasted in SciCast Predict. Data from SciCast isavailable via the SciCast Data Mart [Guide | API].
General Forecasting Tutorials
Most important: Don’t be overconfident! Especially, don’t forecast near the extremes in power mode unless you are very sure, and prepared to lose a lot of points if wrong. (In Safe Mode, you should assert or modify your actual belief.)
SciCast Specific Tutorials
- Why did my forecast do that? — Markets vs Surveys; Power-mode vs. Safe-mode. @DJFlirt3000, this is for you!
- Related Scenarios — Make forecasts depend on the outcome of other questions. Note that these now appear up front instead of afterwards.
- Overview Video (1hr) — Alas, software didn’t capture the live demo at the end.
- Scaled Continuous Question — Instead of estimating the chance of a particular outcome, you are asked to forecast the outcome in natural units.
Video Tutorials
Some Video Tutorials from our old DAGGRE project. The website looks different, and the topic is different, but the fundamentals are the same:
- Basic Strategies — Best 2 minute introduction to forecasting in a market like SciCast
- What is an estimate? — 5 minutes very general overview
- Scoring — 4 min introduction to the core LMSR scoring rule, with examples.
Again, these are from our previous incarnation, so the screenshots won’t look the same, but the core ideas are sound.
Knowledge Base
SciCast Knowledge Base. Searchable by keywords, this is a series of short articles on a variety of topics related to SciCast.
The Official SciCast Blog 