A sailing model connects assumptions about the boat and conditions to a predicted result. Read those assumptions before interpreting a speed, route or percentage. A useful model helps you ask better questions about performance and then test them against appropriate observations.
An explanation for sailors and technically curious readers. The numerical examples are fictional exercises, not navigation or sail-handling instructions.
Separate three different kinds of software
A velocity prediction program estimates boat performance for specified inputs. A data-analysis tool organizes instrument observations. A weather-routing tool combines a performance model with forecasts and route constraints. One application may include all three, but each result answers a different question.
The Offshore Racing Congress VPP documentation describes a physics-based program using hull, appendage, rig and sail information to calculate performance. This is an example of established modeling; the presence of a computer prediction alone does not identify a machine-learning method. For AI-specific design work, see AI in boatbuilding.
Read a polar with its definitions attached
A polar relates predicted boat speed to wind speed and sailing angle. Find the wind-speed curve, then the angle, then read the corresponding boat speed. Check whether the graphic uses true or apparent wind, which sail configuration it represents and which boat condition it assumes. ORC’s Speed Guide explanation shows how its diagrams and performance tables are organized.
On a narrow screen, scroll the table sideways. Keyboard users can focus the table and use the arrow keys.
| Quantity | What it describes | Comparison check |
|---|---|---|
| True wind speed and angle | Wind relative to the stated reference and boat heading | Use the same definition as the polar |
| Apparent wind | Wind experienced by the moving boat | Do not substitute it into a true-wind field |
| Speed through water | Boat motion relative to the surrounding water | Check sensor calibration and model basis |
| Speed over ground | Boat motion relative to the earth | Account for current before comparing with water-relative performance |
Compare like with like on the water
Retain raw logs with timestamps and units. Record boat loading, sails, sea state, current information and calibration status. Choose a reasonably steady interval; flag maneuvering, starting, stopping or obviously invalid sensor readings. Preserve excluded observations and the reason for exclusion so another person can reproduce the comparison.
In the fictional cell above, an observed 6.0 knots through water gives 6.0 ÷ 6.4 × 100 = 93.75%, or about 94% of the prediction. A simultaneous 6.7-knot GPS reading would give about 105%, but it measures a different reference. The difference calls for checking current and sensor definitions before crediting a performance gain.
A persistent mismatch may reflect calibration, loading, hull condition, sea state, sail configuration or a model that fits poorly. Change one test condition at a time where practical and use repeated observations. A single high-speed sample is weak evidence for updating the boat’s performance table.
Understand what a routing result adds
Routing introduces forecast fields, departure time, geography, constraints and an objective such as elapsed time. The Expedition manual emphasizes the importance of the polar, weather and routing settings. A route is therefore conditional on its inputs and settings.
Save the forecast issue time, model, departure time, polar version and limits used. Compare several plausible forecasts or settings to see whether small changes produce a very different route. Check land, restricted areas, depth, actual conditions and operational limits independently through proper navigation practice. The skipper remains responsible for the passage.
Keep a reproducible modeling note
For every useful result, retain the boat configuration, input files, software version, settings, units, result and observed comparison. Label whether a figure is predicted, measured or derived. When sharing a plot, include enough of this context for another sailor to understand what the curve represents.
Begin with one question—for example, whether the observed speed during a steady interval agrees with a particular polar cell. Resolve measurement and definition issues before adding a more complicated model. This gives software output a clear role in learning about the boat.
Use the worksheet
Sailing model and observation worksheet — editable plain text. Save a copy, open it in a text editor and fill in the fields as you collect evidence. Keep unresolved items visible.