A/B Testing On Board: Comparing Two Sail Settings in the Same Wind

Why the "feel" comparison is misleading

Every sailor has lived this scene: change a setting, the boat "feels" faster, but there's no way to prove it. The problem isn't the helmsman's or trimmer's sensitivity — it's that the wind never stays the same for more than a few dozen seconds. Comparing two setups from memory, on different tacks, in different gusts, means comparing different conditions and calling it a "test". Without an objective reading synced to the real wind, any conclusion stays an opinion.

An on-board A/B test applies to sailing the same principle used to compare any two versions of something: change one variable at a time, measure its effect with the same instrument under otherwise equal conditions, and let the data decide, not the moment's impression.

The rules of a comparison you can defend

For a comparison to actually be useful you need a few rules, followed with discipline. The most important is wind parity: two settings should be compared in the same stretch of water, a few minutes apart, not hours apart or on different sides of the racecourse where strength and angle can be very different.

If one of these three conditions is missing, the comparison loses value: you're no longer measuring the effect of the setting, but the effect of the wind that changed in the meantime. With a pressure sensor system like SailSensor Pressure this becomes easier, because every reading is time-stamped and you can overlay the two intervals to see what really changed on the sail, not just on the anemometer.

What to look at in the comparison data

A single peak number isn't enough: an isolated value may depend on a lucky gust. It's worth comparing three levels of reading: average pressure over the interval (shows power generated steadily), the differential between windward and leeward side (shows the profile's efficiency), and peak variability (shows how well the setting holds under gusts). A setting that raises the average but blows up variability can be riskier than one slightly less powerful but more consistent.

If the test concerns boom angle, pairing pressure data with SailSensor BoomAngle lets you read power and angle on the same chart: sometimes the performance difference isn't in the pressure itself, but in how the boom angle distributes it along the mainsail.

The 4-step protocol

Mistakes that invalidate the test

The most common mistake is changing more than one variable at once: if you also touch the forestay tension along with the traveller, you'll no longer know what to attribute the measured difference to. The second mistake is testing in unstable wind, expecting a few seconds of data to be enough to draw conclusions: better to wait for a steadier wind window than force a noisy comparison. The third is confusing the crew with the human variable: if the helmsman or trimmer changes between one interval and the other, execution changes too, not just the setting.

From comparison to decision (and to the boat's memory)

An A/B test only has value if the result stays available for the next outing, not just in the head of whoever was at the helm that day. Saving wind strength, tested setup and result in a log shared with the crew, for example through SailSensor Bridge, turns a single lucky comparison into a knowledge base that grows race after race. In the long run, it's this discipline of comparing settings in the same wind, more than a single flash of genius, that makes the difference on the leaderboard.