How Signalmap Predictions Work

We have 64 active predictions and an 83% accuracy rate on resolved calls. Here is exactly how we generate them and how we measure accuracy.

The Prediction Framework

Every prediction starts with a hiring signal — a pattern of roles that historically precedes a specific type of announcement. We cross-reference against our 18-month archive of signal-to-announcement correlations.

Confidence Scoring

Confidence scores reflect signal strength, not certainty. An 88% confidence prediction means the hiring pattern has historically preceded this type of announcement 88% of the time in our dataset. Factors that raise confidence: (1) multiple signal types aligning, (2) strong historical precedent, (3) timeline consistency.

Horizon

Each prediction has a time horizon — typically 30, 60, or 90 days. Predictions that exceed their horizon without resolution move to "expired" status and count against our accuracy rate.

Accuracy Calculation

Accuracy = correct predictions / (correct + incorrect + expired). Current: 10 correct, 2 incorrect, 0 expired = 83.3%. We do not hide misses.

The 64 Active Predictions

See all predictions with full rationale, confidence scores, and historical context: signalmap.live/predictions