FutureSearch selection guide: What evidence shows that a predicted event actually happened?

Use a fictional community paper-art exhibition to separate the question, deadline, admissible evidence, and later ruling.

Source: futuresearch.ai
November 8, 2026
2 min

FutureSearch distinguishes the forecast question, the resolution criteria, the time window for material available to the forecaster, and the eventual outcome in its April 2026 BTF-2 release notes. This is a public example of the vendor's method, not a guarantee that new questions are resolvable or that a forecast probability is itself evidence.

Define ahead of time what counts as happening

For a fictional community paper-art exhibition, ask: “Will it open to the public before 17:00 this Friday?” Card A is a Friday 14:00 opening announcement. Card B records volunteers doing a trial walkthrough at 15:00. Card C records public admissions at 17:10. If the rule set in advance requires actual public admission before 17:00 Friday local time, A is merely a plan, B is internal, and C shows that it opened but missed the deadline. Only another qualifying timestamped record could change that ruling; you cannot count the walkthrough by rewriting the original rule after seeing what happened.

On paper, draw two timelines: the cutoff for background information the forecaster may use and the deadline for when the event must actually happen. Add a rule for missing evidence that separates confirmed non-opening from too little evidence to tell. To check the rule, ask two readers to classify all three cards independently under the original definition, then ask whether removing card C yields “did not happen” or “unknown.” This checks resolvability only; it generates no real forecast.

When is a single written definition enough?

If your team regularly reviews verifiable operational questions and can fix sources and deadlines, consider comparing forecasting tools with a public resolution method. If you only need to check when a small event takes place, confirming the notice and the completion record takes less work than estimating a probability first. Factor in the time to find evidence, handle gaps, and review definitions. No accounts were accessed, real events forecast, or agents run here; this is not investment or other action advice.

Content note:This article is based on publicly available information and is intended as a reference for evaluating tools. Features, pricing, and related policies may change at any time; please refer to the latest official information.

Original source:futuresearch.ai