Charter Forecast Decision Attribution (CFDA)

GMI-MAR-010Maritime IndustriesIntegration-Required

Definition

Comparative economic performance of commercial decisions that followed ML charter rate forecasts versus those that did not.

The Measurement Gap

What This Metric Measures:
Comparative economic performance of commercial decisions (charter commitments, spot versus period allocation, lay-up timing) that followed ML market forecasts versus decisions that went against forecasts or were made without forecast input. CFDA isolates the ML forecast's contribution to commercial outcomes by comparing decisions grouped by their alignment with the model's recommendations.

The Decision It Supports:

Is the ML charter rate forecasting model producing commercially better decisions than market-conventional judgment, and on which decision types? When the model says spot rates will decline and recommends locking in a period charter, does following that recommendation produce better outcomes than the chartering manager's market instinct?

Why Legacy Metrics Cannot See This:

Fleet commercial performance is measured by TCE, utilization, and period-versus-spot allocation outcomes. None of these metrics attribute performance to the forecast model's contribution. A fleet that follows the ML forecast and earns strong TCE cannot distinguish whether the model added value or whether the market moved favorably regardless. A fleet that ignores the forecast and earns strong TCE does not know what it left on the table. CFDA creates this attribution by comparing outcomes across forecast-aligned, forecast-opposed, and no-forecast decision groups.

Formula

Primary Formula:

CFDA = Mean_outcome_forecast_aligned - Mean_outcome_forecast_opposed

Variable Definitions:

Mean_outcome_forecast_aligned: average net economic outcome (TCE, period rate versus spot index, or net revenue) for commercial decisions that followed the ML forecast recommendation. Dollars.

Mean_outcome_forecast_opposed: average net economic outcome for commercial decisions that went against the ML forecast recommendation. Dollars.

Outcome is measured over the full duration of the decision (charter period, spot voyage, lay-up-to-reactivation cycle).

Units:

Dollars. Positive means forecast-aligned decisions outperformed. Negative means going against the forecast was more profitable.

Edge Cases:

Survivorship bias. If the chartering manager only follows the forecast when it agrees with their judgment, CFDA overstates forecast value. Control for selection bias by tracking the chartering manager's stated intent before seeing the forecast. Market regime changes mean a model that worked in a bull market may fail in a downturn: segment CFDA by market conditions. Minimum sample: 20 commercial decisions per quarter across the fleet.

Proposed formula, pending validation pilots.

At a Glance

Unit
Currency
Trend Direction
Higher; increasing
Reporting Frequency
Quarterly
Category
Voyage and Fleet Commercial Decision Quality
← Survey Certificate Validity Alignment (SCVA)Lay-Up Decision Net Value (LDNV) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment