Context

HomeValueNow built a regression model to predict apartment rental prices for landlords in Chicago. Two candidate models are performing similarly on average, but the pricing team noticed that a small number of large pricing mistakes create outsized business impact because overpriced listings sit vacant and underpriced listings lose revenue.

Current Performance

The Problem

The VP of Pricing asks whether the team should optimize for RMSE or MAE before launch. You need to explain what RMSE measures, how it differs from MAE, and which metric better reflects business risk in this setting.

Requirements

1. Define RMSE and compare it directly with MAE.
2. Interpret the metric differences between Model A and Model B.
3. Explain when RMSE is more appropriate than MAE, and when MAE may be preferred.
4. Recommend which model to launch given the business impact of large errors.
5. Suggest follow-up analyses to validate the recommendation.

Constraints

• Large pricing errors above $500 trigger manual review.
• The business can tolerate small average error but wants to avoid extreme misses.
• Analysts can manually review only 250 listings per week.