Skip to content
ML Academy

Time series and forecasting · insight-0246

Track forecast bias separately from accuracy

Evidence type
practitioner heuristic
Difficulty
easy
Roles
Data analyst, Data scientist
Industries
retail, aviation, logistics

Summary: situation, action, reason, boundary

  1. Situation

    A forecast with acceptable MAE consistently runs below actual demand on high-season weeks.

  2. Action

    Monitor mean error or a cumulative error tracking signal per segment and season alongside accuracy, and alert when bias persists across several periods.

  3. Reason

    Systematic bias is a correctable model or process fault, while random error is not, so they need different responses.

  4. Boundary

    Any operational forecast feeding replenishment, staffing or pricing.

Symptom or risk

Accuracy metrics hide direction; a persistent under-forecast leads to repeated stockouts or underpriced capacity even when average error looks fine.

Exceptions

Intentional bias, such as forecasting a high quantile for service level, should be tracked against the target quantile, not zero.

Trade-offs

Short bias windows trigger false alarms; long windows react slowly.

Pitfalls

Netting positive and negative bias across segments into a reassuring zero.

Lessons and concepts