Update a Forecast With New Evidence
Update a forecast when new evidence is relevant, reliable, and more likely under one outcome than another, while preserving the prior estimate.
Example: delivery forecast
A permit approval raises the chance of on-time delivery from 55 to 70 percent because it removes a known bottleneck. It does not rise to 95 because equipment and staffing risks remain.
Updates should have a trace
Without a recorded prior, people overreact to news or claim they expected the outcome. A simple update log states the old estimate, evidence, direction, size, and reason.
Make a disciplined update
- Record the prior probability.
- Check source and event relevance.
- Ask how expected the evidence was under each outcome.
- Set the new estimate and next trigger.
Repeated weak signals can double count
Several articles may repeat the same source. Trace evidence independence before treating repetition as confirmation.
The decision rule
What to record before acting
Move confidence in proportion to how differently the new evidence fits the competing outcomes.
Apply the rule in context with practice measuring confidence error. Keep the evidence and assumptions with the decision so an outcome review can distinguish reasoning from luck.
Continue within evidence and uncertainty
How to Compare Two Uncertain Timelines
Compare uncertain timelines using milestone ranges, dependency paths, deadline consequences, and the probability each option remains useful when delivered.
How to Compare Two Uncertain TimelinesHow to Use Base Rates in Decisions
Start a forecast with outcomes from a relevant class of similar cases, then adjust for specific differences supported by evidence.
How to Use Base Rates in Decisions