We have entered the age of electric cars, self-driving vehicles, and processors running trillions of operations per second — yet we still most often quote the power of their engines in a unit based on the working strength of a horse from an 18th-century mine. Horsepower is not a mistake of history or a relic nobody got around to removing. It is one of the most successful marketing coups in the history of engineering: a concept invented to sell steam engines that, two and a half centuries on, still resists the official kilowatt.
Horses from the mine gin
The idea of comparing machines to draft animals is older than James Watt. The first to do it was the English inventor Thomas Savery, who in 1702 published a technical-cum-promotional manifesto, The Miner's Friend. Savery wanted to sell mine owners his steam pump and had to convince them to abandon their traditional horse-driven gins. His argument was purely economic: to keep a gin running around the clock, a mine had to keep a dozen or more horses in reserve, resting in the stable, while only two hauled at any given moment. A machine that needs no sleep or oats therefore replaced a whole herd — and it was in horses that Savery priced its worth.
A decade later Thomas Newcomen built the atmospheric engine (1712), far more powerful — one of the first units reportedly replaced the work of hundreds of horses in draining deep shafts. But it was hopelessly inefficient: every cycle required heating and then cooling the same cylinder, devouring vast amounts of coal. The stage was set for someone who could combine efficiency with a brilliant idea for selling it.
Watt and the marketing redefinition of the horse
In 1764 a young instrument-maker at the University of Glasgow, James Watt, was handed a model Newcomen engine to repair. He saw where the energy leaked away and hit on a fix: a separate condensing chamber, which let the cylinder stay permanently hot. His 1769 patent, and the later conversion of reciprocating motion into rotary motion, made his machine the first genuinely efficient source of industrial power.
The real problem arrived after he formed a partnership with the entrepreneur Matthew Boulton in 1775. Watt wanted to sell engines to London breweries and mills — and these had never had any steam engine at all, only horses turning millstones. You could not tell a customer "my engine will save you a third of your coal" when he burned no coal. Watt needed a conversion that would translate pistons and pressures into a language a brewer understood: horses.
So he measured the animals' work. He established that a brewery horse could raise a load of 220 pounds through 100 feet in one minute, which gives 22,000 foot-pounds per minute. Then he did the crucial thing: he arbitrarily raised that figure by half, to 33,000 foot-pounds per minute (that is, 550 foot-pounds per second). The reason was purely commercial — he wanted absolute certainty that no one could ever accuse his machines of being weaker than advertised. An engine sold as "ten horsepower" comfortably beat ten living horses at the gin. Customer satisfaction guaranteed, risk of complaints zero.
The machine's advantage was in fact even greater than the number suggests. A living horse does not work 24 hours a day — a shift ran 8 to 10 hours, so continuous operation needed three shifts of horses. The engine needed no sleep, no hay, no stable. A single "horsepower" running without pause replaced three living animals over a full daily cycle — and with them went the cost of stables, feed stores, and stablehands. All that remained was coal and water.
hp, PS, KM — why there are so many "horses"
Watt defined his horse in imperial units — pounds, feet, and seconds. This is mechanical (imperial) horsepower (hp), still used in the English-speaking world:
1 hp = 550 ft·lbf/s ≈ 745.70 W
When continental Europe went metric, it needed a "horse" based on kilograms and metres. Thus was born metric horsepower (German Pferdestärke, PS; Polish KM) — the power needed to raise 75 kg through 1 metre in 1 second under standard gravity (g = 9.80665 m/s²):
1 PS = 75 kgf·m/s ≈ 735.50 W
Hence a small but misleading confusion: the "horsepower" quoted on most of the world's cars is the metric PS, not Watt's original imperial hp. Both sound like "one horse," yet they differ by about 1.4%. On a small engine that is invisible; at 300 hp it already becomes 304 PS.
The third unit — the only fully scientific one — is the watt (and its multiple, the kilowatt). It belongs to the SI system, rests directly on joules and seconds (1 W = 1 J/s), and depends on no local convention or local gravity. NebulaMath computes every conversion against the watt as its base unit. The table below shows the exact relationships between the most common "horses" and the kilowatt:
| Unit | ≈ W | ≈ kW | ≈ PS | ≈ hp |
|---|---|---|---|---|
| 1 kW (kilowatt) | 1000 | 1 | 1.3596 | 1.3410 |
| 1 PS (metric horsepower) | 735.5 | 0.7355 | 1 | 0.9863 |
| 1 hp (mechanical/imperial) | 745.7 | 0.7457 | 1.0139 | 1 |
| 1 electric hp | 746 | 0.746 | 1.0143 | 1.0004 |
| 1 boiler hp | 9810 | 9.81 | 13.34 | 13.15 |
The practical takeaway: a figure in horses is always about a third larger than the same power in kilowatts. 100 kW is nearly 136 PS — which is exactly why marketing loves horses. A "bigger number" sounds more powerful, even when it describes precisely the same power.
Torque is not power
Since we are talking about engines, it is worth separating two things that adverts blur into one: power and torque. An engine's power (P, in kilowatts) follows directly from its torque (T, in newton-metres) and its rotational speed (n, in revolutions per minute):
P = (T · n) / 9549.3 — power in kW is torque in N·m times rpm, divided by the constant 9549.3.
Why can two engines with identical peak power drive completely differently? Because torque decides pulling force — how eagerly the car moves off the line — while power decides how high a top speed it can reach.
The best image is two labourers shifting the same heap of soil in the same time. The first — the embodiment of torque — has a huge shovel: he scoops enormous loads and moves the lot in a few slow, calm strokes. The second — the embodiment of power — has a tiny trowel and must swing it very fast, at high "revs," to keep up. The work done is identical, yet the character of the effort is utterly different. That is exactly how a lazy diesel with massive low-end torque differs from a high-revving engine that only wakes up near the redline.
How many horsepower does a horse have?
Here lies the best paradox of the whole story: one horsepower is far less than the power a real horse can produce. It all depends on whether we mean average power (sustained for hours) or peak power (in a burst of a few seconds).
In continuous work, over a full day, a healthy draft horse produces on average 0.7–0.8 PS — less than the unit named after it. Watt's original estimates were therefore surprisingly accurate, and his marketing inflation of the standard to "1 PS" pushed the reference point above the real, sustainable capacity of a single animal.
An explosive burst is a completely different matter. In a dynamic start, a horse's muscles (about 45% of its mass) can produce enormous instantaneous power. The classic study by biologists R. D. Stevenson and R. J. Wassersug, published in Nature in 1993, estimated a horse's peak power at about 14.9 hp — more than ten times its steady working output. A human comes off modestly by comparison, but not hopelessly: in a full day of physical work we sustain roughly 100 W (about 0.13 PS), yet in a short burst an elite sprinter can reach 2,200 W — close to 3 PS.
| Subject | Type of effort | Power |
|---|---|---|
| Draft horse | continuous (8–10 h a day) | 0.7–0.8 PS (≈ 515–590 W) |
| Draft horse | peak burst (a few s) | 12–15 PS (≈ 9–11 kW) |
| Average human | sustained (working day) | ≈ 0.13 PS (≈ 100 W) |
| Average human | short anaerobic burst | 1.2–1.5 PS (≈ 900–1100 W) |
| Road cyclist (PRO) | one hour (FTP) | ≈ 0.55–0.68 PS (400–500 W) |
| Track sprinter | finish (up to ~5 s) | 2–3 PS (1500–2200 W) |
A curiosity: a professional cyclist's one-hour threshold power (FTP, 400–500 W) is almost what an adult draft horse pulls all day. One trained human matches a horse for an hour — except the horse does it from dawn to dusk.
Why the authorities demand kilowatts
In drivers' minds horsepower reigns, but officialdom has long counted differently. In the European Union, Directive 80/181/EEC on units of measurement makes the kilowatt the only fully legal, official unit of engine power. Horsepower (PS, hp) is permitted solely as a supplementary indication — a maker may print it in an advert or brochure, but always alongside the dominant value in kW. In Poland the primacy of the kilowatt is anchored in the Law on Measures.
That is why the vehicle registration document has no room for horsepower: engine power goes only into the metric field P.2 (maximum net power in kW). This has practical consequences when importing a car from, say, the US, where the papers state power in hp — for the entry to pass validation against the type-approval database, the value must first be converted to kilowatts with the proper factor (0.7457 for hp).
The kilowatt can hit the wallet too. Insurers pull a car's power from their databases in kW, but an insurance calculator makes a costly slip easy. Someone thinking in "horses" who types "150" into a field labelled kW (meaning a 150-horsepower car whose actual power is about 110 kW) is in fact declaring a car of over 200 hp. The system assigns the vehicle to a higher risk band, and the premium can then jump by tens of percent. It pays to watch which unit the form is asking for.
The "system power" of hybrids has become its own battleground. For years makers simply added the peak power of the combustion engine and the electric motor arithmetically (e.g. 150 PS + 100 PS = 250 PS). Physically this is an abuse: the peak outputs of the two occur at different speeds, and the whole is limited anyway by the battery, which can rarely feed both motors at full power at once. The real system power is lower — and today regulators and consumer-protection bodies increasingly push for it to be stated in type-approved kilowatts rather than as a flashy, summed horsepower figure.
The takeaway: the slow arrival of the kilowatt
Horsepower's longevity comes not from any scientific usefulness — it has none — but from psychology. Watt brilliantly sensed that people need a "cognitive bridge" between the familiar world of horses and the incomprehensible Industrial Revolution. Horsepower demystified the steam engine, giving it a human, imaginable scale. And that the resulting number is a third larger than in kilowatts — all the better for sales.
That era, however, is slowly ending. In electric cars the natural currency has become the kilowatt and the kilowatt-hour: we quote battery capacity in kWh, charging rate and motor power in kW. The more such cars on the road, the more the kilowatt stops being an "official abstraction" and becomes a number the driver feels intuitively. Until then, horsepower will remain a fascinating monument to 18th-century marketing — the thread that ties today's supercars to a horse walking in circles in James Watt's mine gin.
Further reading
- R. D. Stevenson, R. J. Wassersug, Horsepower from a horse, Nature 364, 195 (1993) — the source of a real horse's peak ~14.9 hp.
- Encyclopædia Britannica, entry James Watt — the biography and the origin of the Boulton partnership.
- BIPM, The International System of Units (SI Brochure), 9th ed. (2019) — the definition of the watt and the place of power in SI.
- Council Directive 80/181/EEC on units of measurement — the kilowatt as the sole legal unit of power in the EU, horsepower as a supplementary indication.