PolarForge

Introduction

PolarForge takes raw sailing log data and turns it into an actionable polar diagram. Upload one or more Expedition CSV log files, compare recorded performance against a baseline polar, accept the cells where you beat it, and export the updated polar for use in navigation software.

The app ships with the F35 ORC 2022 polar pre-loaded as a demo baseline so you can explore the interface immediately without any data.

Workflow at a glance

Upload Performance Logs

PolarForge reads Expedition CSV log files — the standard export format from the Expedition navigation and routing application.

How to upload

Either drag and drop one or more .csv files onto the upload zone in the Performance Logs panel (left sidebar), or click the zone to open a file browser. You can also use the + Upload Logs button in the header. Multiple files are merged into a single dataset.

Required CSV columns

After uploading

When a file includes a Utc timestamp column, a timeline preview modal opens automatically before the data is imported. This lets you select only the racing segments you want. See Timeline Selection for details.

Import a Baseline Polar

A baseline polar gives PolarForge a reference to measure your recorded data against. Without one, the recorded data is still displayed but no improvements can be detected.

Click ⬆ Import Polar in the header, then paste polar data into the text box or upload a file (.pol, .txt, .csv, or .json). Two formats are supported:

Expedition format

Tab-separated. Each data row starts with a TWS value followed by TWA BSP pairs.

!Expedition polar
6.0	43.2	4.64	52.0	5.18	60.0	5.49	90.0	5.82
8.0	41.8	5.50	52.0	6.15	60.0	6.45	90.0	6.70
10.0	40.5	6.20	52.0	6.95	60.0	7.25	90.0	7.60

Grid CSV format

TWA values in the first column, TWS values as column headers, BSP values in the body.

TWA,6,8,10,12,16
30,3.2,4.1,5.0,5.5,6.0
45,4.5,5.5,6.2,6.8,7.4
60,5.2,6.3,7.1,7.7,8.3
90,5.8,6.9,7.8,8.4,9.0

JSON format

Nested object exported by PolarForge itself. Keys are TWS values; inner keys are TWA angles; values are BSP.

{ "6": { "43": 4.64, "52": 5.18 }, "8": { "41": 5.50, "52": 6.15 } }

Reading the Polar Chart

The polar diagram is a half-circle speed plot. Each point on a curve represents the boat speed achievable at a given wind angle for a specific true wind speed. The further from the centre, the faster the boat.

Axes

Legend — what the colours mean

Navigation

Visualization Settings

The ⚙ Settings panel (right column) controls how data is aggregated and what is shown on the chart.

Direct Point Editing on the Chart

Individual data points on the polar diagram can be edited directly — no need to find the cell in the table. Hold the left mouse button on any curve point for 2 seconds to enter edit mode, then use the scroll wheel to adjust the value. The curve updates live as you scroll.

How to use it

The edit popup

What updates live

While scrolling, the chart redraws the affected wind speed curve in real time. The gold dot tracks the new position along the radial line. The popup delta updates with each notch. Accept triggers a full refresh: improvements are re-evaluated, the VMG panel updates, and the table reflects the new value.

Data Filters

All uploaded CSV data passes through the Data Filters panel before it is added to the polar. Filters remove unreliable or non-sailing data points. All active filters are combined with AND logic — a point must pass every filter to be included.

Wind Speed Filter

The Wind Speed Filter panel shows a button for every TWS band present in your data. Clicking a band toggles its visibility on the polar chart. This lets you focus on specific wind conditions without deleting any data.

Toggle All shows or hides every band at once.

Timeline Selection

When an uploaded CSV file contains a Utc timestamp column, a timeline preview modal opens before the data is imported. This lets you select only the portions of the log you want — typically, the clean upwind and downwind legs of a race, excluding start-line chaos, mark roundings, and any time spent motoring.

Reading the timeline chart

The chart shows the full session as a multi-line overlay:

Selecting segments

Improvements

After importing data, PolarForge compares every (TWS, TWA) cell in the recorded polar against the corresponding cell in the baseline. Cells where your recorded speed exceeds the baseline are called improvements and are shown in green on the chart.

Improvement banner

When improvements are found, a banner appears at the top of the page showing a summary: number of improved cells per wind speed and the average percentage gain. Click ☰ Details to expand the full list.

Reading the details panel

Improvements are grouped by wind speed. For each cell you'll see:

Applying improvements

Once you are satisfied that an improvement is genuine (not noise), apply it to merge it into the base polar:

Confidence scaling

Larger claimed improvements require proportionally more supporting samples. A cell claiming a 10% gain needs significantly more data points before it is accepted than one claiming a 1% gain. This prevents a handful of lucky fast runs from inflating your polar.

Polar Data Table

Below the polar chart, a data table displays all polar values in a grid with TWA angles as rows and TWS bands as columns. It is the main place to review and understand what is in your polar.

Table views

Three views are available via buttons above the table:

Editing Cells

Any cell in the Speed view can be edited manually. Click a cell to enter edit mode; the cell turns into an input field.

Enter a positive number to set a new BSP, or leave the field blank to clear the cell.

Editing column headers

Double-click a TWS column header to rename the wind speed value. Hover a header to reveal a × Delete button that removes the entire column.

Adding rows and columns

VMG Optimal Angles

The VMG Optimal Angles panel shows the best angle and speed for each wind speed band in your polar. VMG (Velocity Made Good) is the component of boat speed that is actually making progress toward the target — directly upwind or downwind.

For each TWS the panel shows two rows:

VMG is calculated as: BSP × cos(TWA) for upwind and BSP × cos(180° − TWA) for downwind.

Session Statistics

The Session Stats panel (appears when log data is loaded) shows a quick summary of your imported data:

Compare Polars

The Compare Polar panel lets you load a second polar (Polar B) to compare side-by-side with your base polar (Polar A).

Polar B is also rendered as a purple curve in the diagram.

Fill angle extremes

If Polar B has data at extreme angles (very low or very high TWA) that your base polar lacks, clicking Fill angle extremes copies those cells from B into A. This is useful when extending a manually trimmed polar with data from a more complete source.

Clear

Click Clear in the Compare Polar panel to remove Polar B entirely.

Exporting Data

Three export buttons are available in the header. All exports use the merged polar (baseline + any applied improvements).

Key Concepts

True Wind Speed (TWS)

The actual wind speed, independent of the boat's motion. This is the primary grouping dimension for polar data — performance is compared only within the same wind speed band.

True Wind Angle (TWA)

The angle between the wind direction and the boat's heading, measured from 0° (directly into the wind) to 180° (directly downwind). Pointing 45° off the wind is a close-hauled upwind angle; 150° is a broad reach or deep downwind angle.

Boat Speed (BSP)

Speed through water, not affected by current. This is the key performance metric. Speed Over Ground (SOG) includes current effects and is used only for filtering.

Baseline polar

The reference curve — usually an ORC or IMS rating polar, a manufacturer specification, or a previous season's best. All improvements are measured relative to this.

Merged polar

The final combined polar: baseline values are used where no improvement exists; improved values replace the baseline where recorded data is better. This is what you export.

Aggregation method

When multiple logged data points fall in the same (TWS, TWA) bin, they are combined by the selected aggregation method. The 95th percentile is recommended for racing polars — it represents a genuinely fast run rather than the average, while still being robust against the occasional freak reading.

Tips & Tricks

Getting clean data

• Use the timeline to select only steady sailing legs — long upwind or downwind stretches produce the best signal.
• Set a minimum heel angle of 3–5° to exclude any motoring or drifting data.
• The After gun race mode filter automatically excludes pre-start maneuvers when the CSV contains a TmToGun column.

Building confidence

• Switch to the Coverage view before applying improvements — a cell with fewer than 20 samples is unlikely to be reliable.
• Raise min samples per cell to 5–10 if you have a lot of data; this filters out bins with very few observations.
• Upload logs from multiple races in the same wind range to accumulate samples across the same bins.

Refining the polar

• Apply improvements wind speed by wind speed. Carefully review any cell claiming a gain of more than 5% — that level of improvement warrants more data.
• Edit individual cells manually if one value looks physically impossible (e.g. faster than hull speed).
• Enable Smooth improvements to interpolate naturally between the improved cells.

Version management

• Export a JSON snapshot at the end of each session.
• Load a previous snapshot as Polar B to quantify how much the polar has improved since last time.

Troubleshooting