How to Test Your Boat's Real Speed with GPS

Introduction

Your knotmeter says 6.2 kn (7.1 mph)6.2 kn (11.5 km/h). The phone clipped to the helm says 7.4 kn (8.5 mph)7.4 kn (13.7 km/h). Same boat, same second, two numbers more than a knot apart. Which one is lying?

Neither. Your knotmeter measures how fast water is flowing past the hull; GPS measures how fast the hull is moving across the seabed. The gap between them is the current you are sitting in — tangled up with whatever your knotmeter is getting wrong. Separating those two is what a boat speed test is for, and it takes about ten minutes.

Key Takeaways

• Your knotmeter measures speed through the water; GPS measures speed over the ground. On a course straight up- or down-current, the difference between them is the current plus any error in the knotmeter.
• A single pass is worthless for measuring your boat (it is a perfectly good speed over the ground for navigation). Run the same course both ways and average the two — a steady current cancels itself out exactly.
• The two-way average is your speed through the water. That, not a single GPS pass, is what you calibrate the knotmeter against; calibrating against one pass bakes the current into the instrument.
• Average speed is the honest number. A maximum is a one-second spike — it tells you about a wave, not about your boat.
• 1 knot = 1 nautical mile per hour = 1.15078 mph1.852 km/h, and a nautical mile is 1,852 metres (6,076 feet)1,852 metres exactly by definition.

Why Your Boat’s Own Instruments Drift

Marine speed instruments are not bad engineering. They are old engineering, and every one of them measures something that stands in for speed rather than speed itself.

The paddle wheel counts turns, not progress. The most common log on production boats is a small paddle wheel in a through-hull fitting. Water spins it, a magnet triggers a pickup, and the display converts pulses into knots using a constant burned in at the factory — a constant that assumes the wheel is clean, the bearing free and the flow undisturbed. All three decay. A season of slime and barnacle growth adds drag, the wheel turns slower than the water around it, and the instrument under-reads. It is the same failure mode as a car counting wheel rotations and assuming a tire diameter, which is exactly why GPS speed is more trustworthy than a vehicle’s own gauge.

Pitot tubes measure pressure, not speed. Fast planing boats often use a forward-facing port that senses the pressure of water forced into it. Pressure rises with the square of speed, so the gauge is responsive at the top of the range and nearly useless at the bottom: below roughly 8–10 kn (9–12 mph)8–10 kn (15–19 km/h) many read low, read erratically, or do not register at all. They also block easily — a wasp nest over the winter, a fleck of weed — and a partial blockage bleeds off sensed pressure, so the gauge quietly reads low with nothing obviously broken.

Where the transducer sits matters. A speed transducer samples the layer of water your hull has already disturbed, not the open sea. Worse, the error changes with trim: a boat at rest sits level, the same boat on the plane sits bow-up with a different section of hull wetted. That is why a knotmeter can be reasonably accurate at displacement speed and badly wrong at cruise, and why a single-point calibration rarely fixes a boat across its whole range.

None of this is fixable from the helm. What is fixable is having a second, independent speed source to check against, which is what a phone running a GPS boat speedometer provides: no through-hull, no moving part, nothing to foul.

Speed Through Water vs Speed Over Ground

There are two legitimate answers to “how fast is this boat going”, and they are different quantities.

Speed through the water (STW) is your speed relative to the water you are floating in — what a paddle wheel or pitot tube measures, and the number that matters for sail trim and hull performance. Speed over the ground (SOG) is your speed relative to the seabed — what GPS measures, and the number that matters for passage planning, your ETA, and whether you will clear that headland before the tide turns.

The difference between the two is the current. That is not a rule of thumb, it is vector arithmetic: your motion over the ground is your motion through the water plus the motion of the water itself.

Go back to the pair of numbers from the introduction. The knotmeter reads 6.2 kn (7.1 mph)6.2 kn (11.5 km/h) and GPS reads 7.4 kn (8.5 mph)7.4 kn (13.7 km/h). On a straight course with the current directly behind you, that gap of 1.2 kn (1.4 mph)1.2 kn (2.2 km/h) is the tide pushing you along — plus whatever your knotmeter is getting wrong. One reading cannot separate the two. Turn around and run the reciprocal course, and the same tide now subtracts: the knotmeter still reads about 6.2, but GPS drops to roughly 5.6 kn (6.4 mph)5.6 kn (10.4 km/h).

Nothing broke between those two runs. The boat did not slow down — you stopped being helped and started being hindered, and only the GPS number noticed, because only it is referenced to something that does not move. The two agree when the water is still and the knotmeter is honest: on a calm lake they should match, on a tidal river they can differ by several knots and both be right.

What a GPS Speed Test Actually Measures

A GPS speedometer does not take two position fixes and divide by the time. It derives velocity from the Doppler shift on the satellite carrier signal: your motion relative to each satellite stretches or compresses the received frequency by a tiny, measurable amount, and with four or more satellites in view the receiver solves those shifts for a three-dimensional velocity vector. Speed over the ground is its horizontal component.

Two things follow. It is fast, because a frequency measurement needs no long baseline — the reading settles within a second or two of a speed change rather than lagging it. And it is precise at speed: position accuracy and velocity accuracy are separate specifications, so your position can wander a few metres while the velocity solution stays good to a few tenths of a knot with a clear view of the sky. The mechanics are covered properly in how GPS speedometers work.

What GPS cannot tell you is your speed through the water. It has no way to know the whole ocean is drifting at two knots — from a satellite’s point of view, water is scenery. That is not a limitation to work around; it is exactly the property that makes GPS the reference you check everything else against.

The Two-Way Run: A Boat Speed Test That Works

If you take one GPS reading on one heading and call it your boat’s speed, you have measured your boat plus whatever the water was doing. On a tidal coast that can be a two-knot lie in either direction.

The fix is the two-way run, borrowed from the measured-mile trials that boatbuilders have used for a century. You run the same course in both directions and average the two results. A steady current adds to one pass and subtracts from the other by the same amount, so the average is your speed with the current removed.

How to run it

1. Pick water you can use. You want a straight, unobstructed stretch, deep enough that you are not fighting shallow-water drag, out of the shipping channel, and free of traffic that will make you alter course mid-run. Depth matters more than people expect, and not in one direction: shallow water slows a displacement hull, while a planing hull can actually run faster over a shallow bottom. Either way depth becomes a variable you did not intend to measure, and the planing case is the dangerous one because it flatters the result.
2. Pick a length. At least 0.5 NM (0.58 statute miles)0.5 NM (0.93 km) per pass, and 1 NM (1.15 statute miles)1 NM (1.85 km) if you have the room. Longer runs average out wave-to-wave variation. Shorter runs are dominated by the moment you happened to sample.
3. Settle the boat first. Come up to the throttle setting you want to test, hold it, and give the boat time to reach a steady attitude and speed before the run starts. On a planing hull that can take fifteen or twenty seconds. Trim, tabs and load should be exactly where you intend to leave them.
4. Run the first pass on a fixed heading, holding the throttle constant, steering as straight as the conditions allow. Do not chase the speed number with the throttle. Record the average GPS speed for the pass.
5. Turn around and run the reciprocal, same throttle, same trim, same water, opposite direction. Turn wide enough that you are back at a settled speed before the second pass starts.
6. Do it again if the numbers look odd. Two pairs of passes cost five extra minutes and will tell you immediately whether the first pair was contaminated.

The arithmetic

Take the two averages and add them together, then divide by two. That is all.

Say the downstream pass averaged 7.4 kn (8.5 mph)7.4 kn (13.7 km/h) and the upstream pass averaged 5.6 kn (6.4 mph)5.6 kn (10.4 km/h):

true speed  = (7.4 + 5.6) / 2 = 6.5 kn
current     = (7.4 − 5.6) / 2 = 0.9 kn

Your boat is doing 6.5 kn (7.5 mph)6.5 kn (12.0 km/h) through the water at that throttle setting, and you were sitting in a 0.9 kn (1.0 mph)0.9 kn (1.7 km/h) current. Both numbers came out of the same two passes.

The arithmetic mean cancels a constant, along-course current exactly. That is the assumption the whole method rests on, and it has three ways of failing.

A turning tide. If the current changes while you run, the average is biased by however much it moved between passes — which is why the passes go back to back, not an hour apart over lunch. Running two pairs and comparing them detects this, but it does not correct it. If you want the correction, the measured-mile answer is four runs weighted 1:3:3:1 and divided by eight, which cancels a steadily changing current exactly:

true speed = (run1 + 3 × run2 + 3 × run3 + run4) / 8

A current across your course. The method assumes the water is moving along your track, not across it. A pure beam current is the nasty case: it slows you by the same amount on both passes, so the two readings agree, the maths reports zero current, and the average comes out slightly high. It looks like a clean result and it is not. Pick a course aligned with the tidal stream, which on most coasts means along the channel rather than across it.

A boat that is not doing the same thing twice. Wind and sea state change your speed through the water, so if the wind gets up between passes, the two runs are not measuring the same boat.

Wind is a different case, and worth being clear about. Wind pushes on your rig and topsides and changes your speed through the water, so a headwind genuinely slows the boat and the two-way average will be lower than on a calm day. What wind does not do is bias the GPS reading. The receiver reports how fast the hull is moving over the seabed, whatever combination of thrust, current and windage produced it.

What to Record

Most speed apps will offer you a maximum, an average and a distance. They are not equally useful, and treating them as interchangeable is how people end up quoting a top speed their boat has never actually held.

Maximum speed is the highest single reading in the run — the number everyone wants and the one that means least. It is a one-second event: the boat surfing down the face of a wave, or catching a moment of favourable set. Fine for bragging, useless for comparison.

Average speed is the honest figure — what the boat actually sustained over the whole pass, and the only one of the three that compares meaningfully between runs, throttle settings or seasons. It is also what the two-way run is built on: you average the average of each pass, never the peaks.

Distance is your cross-check. If the app says you covered 0.94 NM (1.08 statute miles)0.94 NM (1.74 km) on a run you intended to be 1 NM (1.15 statute miles)1 NM (1.85 km), the pass was cut short and the average is describing a shorter stretch of water than you planned.

Write down the conditions alongside the numbers: throttle setting or engine RPM, trim and tab position, fuel and water aboard, number of people, sea state, and wind direction relative to the course. A boat that made 6.5 kn (7.5 mph)6.5 kn (12.0 km/h) light and flat will not repeat it with full tanks and four guests, and if you did not note the difference you will spend the winter wondering what went wrong with the engine.

This is why the app you use should log a trip rather than just display a needle: recording the run gives you an average over a defined stretch instead of a number you squinted at while steering. A GPS speedometer set to knots holds max, average and distance per trip — exactly the three-column record this method wants.

Why Boat Speed Is Measured in Knots

A knot is one nautical mile per hour, and a nautical mile is 1,852 metres (6,076 feet)1,852 metres exactly by definition — a figure chosen to sit very close to one minute of latitude. That gives it a property no statute mile or kilometre has: the latitude scale up the side of a nautical chart doubles as a distance scale, so a boat doing 10 knots covers ten minutes of latitude in an hour. Step the dividers off against the part of that scale abeam your position, not the top or bottom edge of the sheet — on a Mercator chart the scale stretches as you go north or south.

The conversions, for when you need to talk to someone on land:

Calibrating Your Knotmeter From the GPS Result

The two-way run gives you the correction factor for free. You already have the true speed through the water; all you need is what the knotmeter said while you were doing it.

From the worked example: the two-way average was 6.5 kn (7.5 mph)6.5 kn (12.0 km/h), and the knotmeter read 6.2 kn (7.1 mph)6.2 kn (11.5 km/h) on both passes. The instrument is under-reading:

correction = 6.5 / 6.2 = 1.048  →  enter +4.8% (the knotmeter reads low)

Most modern instrument displays expose a boat-speed calibration setting, either as a single percentage or as a multi-point table where you enter a correction at several speeds. Apply the factor, then run the test again to confirm it took. Three things worth knowing before you do:

Clean the transducer first. There is no point calibrating out fouling. Pull the paddle wheel, clear the growth, check it spins freely, then measure — otherwise you bake this season’s barnacles into the instrument’s settings and have to redo it when they fall off.

Calibrate at more than one speed if your system allows it. Paddle wheels are not linear. A wheel that reads 5% low at displacement speed may read correctly at cruise, or worse, so a single-point correction fixes one speed and shifts the error elsewhere. If your display only offers one number, calibrate at the speed you actually spend your time at.

Redo it once a season. Growth comes back and bearings wear; the correction that was right in May will not be right in September.

If the knotmeter is off by more than about 15% after cleaning, or its error changes wildly with speed, calibration is papering over a hardware problem — a damaged paddle, a failing pickup, or a transducer in bad flow. That is a haul-out job, not a menu setting.

Frequently Asked Questions

How do I check my boat speed in knots right now?

Any GPS speedometer set to knots will show it. GPS gives you speed over the ground directly, with no through-hull fitting, no calibration and nothing to foul, so a phone at the helm needs no setup beyond a clear view of the sky and a few seconds to get its first fix. Bear in mind that this is your speed relative to the seabed, so in a current it will differ from your boat’s own knotmeter — that difference is the current, not a fault in either device.

Why does my knotmeter disagree with my phone’s GPS?

Because they measure different things. The knotmeter measures speed through the water; GPS measures speed over the ground. Any current sets them apart. On a course run straight into or with the current, the gap is the current plus any knotmeter error — which is why one reading cannot give you either number on its own. If the two still disagree in genuinely still water — a lake on a calm day — then you have a real instrument error, most often paddle-wheel fouling, and the two-way run will tell you how big it is.

How accurate is a GPS boat speed test?

At speed and with a clear view of the sky, a GPS velocity solution is typically good to a few tenths of a knot, because velocity comes from the Doppler shift on the satellite signal rather than from differencing positions. That is far tighter than a paddle wheel or a pitot tube. Speed is the magnitude of a noisy velocity vector, so at rest that noise cannot cancel to zero — it rectifies into a small positive number. A GPS reading of “0.3 knots” while tied to a dock is that artefact, not movement.

How long should a boat speed test course be, and how many passes?

At least 0.5 NM (0.58 statute miles)0.5 NM (0.93 km) per pass, ideally 1 NM (1.15 statute miles)1 NM (1.85 km), and always two passes on reciprocal headings run back to back. One pass in each direction is the minimum that lets the current cancel; two pairs is better, because agreement between the pairs tells you the water was steady while you tested.

Do wind and current affect the GPS speed reading?

They affect your boat, not the reading. A current physically moves your hull over the ground, so GPS correctly shows you going faster or slower — that is the measurement working, not failing. Wind changes your speed through the water by pushing on the rig and topsides, which GPS also correctly reports. Neither introduces an error into the GPS figure itself; averaging two reciprocal passes removes the current’s contribution and leaves your boat’s own performance.

Sources & further reading

• BIPM — SI Brochure, Table 8 (non-SI units) — the international nautical mile as exactly 1,852 metres.
• GPS Standard Positioning Service performance (gps.gov) — the official civilian GPS accuracy specification, including the velocity figures behind the Doppler claims above.
• NOAA — Nautical chart use and the latitude scale — why distance is stepped off against the latitude scale abeam your position.

Conclusion

The knotmeter and the phone were never in competition. One tells you how the hull is moving through the water, which is what you steer and trim by. The other tells you how you are moving across the chart, which is what gets you home before the tide turns. A boat that knows both numbers knows the current for free.

A boat speed check takes one straight stretch of deep water: settle at the throttle setting you care about, run half a nautical mile, turn around, run it back, and average the two. Write down the average, not the maximum, and note the conditions beside it. Do it once a season and you will spot a fouled paddle wheel from the helm instead of discovering it on the hard.

If you want the marine setup — knots as the default unit, max, average and distance logged per trip, and a display you can read at the helm — the GPS boat speedometer page covers how it works on the water.

Download GPS Speedometer and run your first two-way pass this weekend. Core speed tracking works offline and displays knots, mph and km/h; optional PRO features are available.